fix(bin): emit a real @AGENTS.md pointer instead of a CLAUDE.md symlink - #2512
Merged
Conversation
…mlink A Write aimed at CLAUDE.md followed the symlink and destroyed AGENTS.md. The installer now creates and migrates to a recoverable two-line pointer file.
This was referenced Aug 17, 2026
AmadeuReal
pushed a commit
to AmadeuReal/firstmate
that referenced
this pull request
Aug 17, 2026
…mlink (kunchenguid#2512) A Write aimed at CLAUDE.md followed the symlink and destroyed AGENTS.md. The installer now creates and migrates to a recoverable two-line pointer file.
EvanAgee
added a commit
to EvanAgee/firstmate
that referenced
this pull request
Aug 17, 2026
* fix(decisions): close decision holds at answer time via one general keyed-answer path (kunchenguid#2490) * fix(decisions): close captain holds at answer time Firstmate had two "a decision is open" ledgers with asymmetric closing mechanics. The live status-log ledger closes atomically at answer time, because bin/fm-send.sh --resolve-key makes answering a decision be the act that closes it. The durable backlog hold ledger had no such coupling: answering and recording were two separate acts, and only the first was forced by the workflow. That asymmetry lost four real captain decisions. Their answers were captured durably to disk, keyed character for character by the hold decision keys, acknowledged, and even implemented and shipped, yet the holds stayed open for two days and the captain was asked to re-answer decisions already on his own disk. Give the hold ledger the same answer-time-closure property: - bin/fm-decision-hold.sh gains an `answer` subcommand, the hold ledger's counterpart to --resolve-key. It shares one unrouted close implementation with `decline`, so it carries every existing guard - the captain decision file, the active-hold requirement, retry identity, and the refusal to release still-routed work - and differs only in the resolution mode it records. `decline` keeps its stronger meaning that the answer routes no follow-up work at all. - bin/fm-procevent-lavish.sh wires the channel that actually carried the lost answers. `arm --decisions-origin` binds a deck to the origin whose holds it carries, `answers` reads the structured choices out of a captured poll result, `close-decisions` maps each key to its hold and closes it through the command above, and `autohandle` lets the runner apply that at capture time. Safety is preserved rather than traded away. Only rows tagged `choice` are read, so freeform captain prose cannot forge a decision key. Closure is confined to the one bound origin. The decision text is a pure function of the captured result, so a replayed capture is idempotent. A hold that is absent, already closed, or still blocking routed work is skipped and left for `resolve`, never forced. A deck armed without the binding touches no hold at all. And autohandle deliberately never reports full handling, because recording an answer is transcription while acting on it is firstmate's judgement - so the check wake still reaches the handler. fm-send --resolve-key is untouched. * no-mistakes(document): document state/lavish-decisions binding dir in AGENTS.md state inventory * refactor(decisions): make keyed-answer closure one general capability The previous pass gave holds answer-time closure but built it as bespoke Lavish wiring: the review adapter carried the source-to-origin binding, mapped keys to hold identities, wrote decision records, decided what to skip, and closed holds itself. That treated a review deck as a special decision source. It is not - it is an ephemeral discussion format that happens to carry answers. Collapse it into ONE general capability with one owner. bin/fm-decision-hold.sh now owns the whole of "a keyed answer closes its matching hold": - `answers <origin> --source <provenance>` is the channel-agnostic intake. It reads key/answer/label lines on stdin, maps each key to its hold, and closes it through the same `answer` path, so every guard applies identically whatever channel the answer came from. --source is provenance recorded in the decision, never a behavior switch; there is no per-channel branch and no knowledge of chat, decks, or transports. - `bind`/`unbind`/`binding` own the source-to-origin binding for any channel whose answers arrive detached from their origin. Every channel is now an ordinary caller that only turns what it received into keyed lines: - bin/fm-send.sh (chat) feeds the intake for a key that names an active hold. This also fixes a real gap: once `complete` transfers a decision to its hold it closes the live status copy, so --resolve-key alone could never answer a transferred decision. - bin/fm-procevent.sh feeds it generically. A bound source's captured result goes to `<adapter> answers <result-file>` and whatever that prints is piped into the intake. The runner names no adapter, parses no result, and carries no decision rule, so any future adapter with an `answers` command works with no change here. - bin/fm-procevent-lavish.sh keeps only `answers`, which reports the structured choices a review captured and stops. It maps nothing to a hold and closes nothing; it lost ~160 lines of decision logic. Feeding is independent of handling, so it never acknowledges a result and never suppresses a wake - recording an answer is transcription, acting on it stays firstmate's judgement. The regression that proves closure now drives a FIXTURE adapter that is not the review adapter, so what is proven is that any bound channel reaches the intake rather than that one channel is wired specially. A new regression drives the real fm-send over a stubbed transport for the chat side. Every prior guarantee still holds, and fm-send's status-log behavior is unchanged. * no-mistakes(review): test(decisions): drop source-content grep from hold-closure regression * fix(memory): emit a real @AGENTS.md pointer instead of a CLAUDE.md symlink (kunchenguid#2512) A Write aimed at CLAUDE.md followed the symlink and destroyed AGENTS.md. The installer now creates and migrates to a recoverable two-line pointer file. * fix(ci): keep CLAUDE.md pointer check valid (kunchenguid#2515) * feat(outage): keep landing work while GitHub is down Two coupled parts so the fleet keeps moving during a GitHub outage. Part A - detection: - bin/fm-github-health.sh: a bounded, mostly-unauthenticated GitHub reachability probe. Silent when reachable; on confirmed unreachability (majority of attempts fail, so one bad request never parks the fleet) it sets state/.github-down with a first-seen timestamp, and clears it on recovery. A transition command owns the report-once-per-change dedup. - bin/fm-watch.sh: calls the probe once per check interval and wakes the captain once on each down/up change, never per poll. Part B - local landing and sync-on-return: - bin/fm-merge-local.sh: generalized to take an explicit lane branch (or read it from the task worktree) and to let a normally-PR-bound task land locally only while GitHub is down. Every existing guard is kept: clean tree, default branch, fast-forward-only, diverged-refusal. A yolo-on outage auto-land also requires proof that a second, independent agent adversarially reviewed the change (the captain's decision), so an auto-land can never happen on one review. Each outage landing is recorded in state/outage-landings/<project>.log. - bin/fm-outage-sync.sh: on return, per ledger entry it fast-forward-pushes local main when clean-ahead, dispatches the deferred schedule-only workflows, and clears the entry. A diverged main escalates rather than force-pushing. Idempotent: a commit already on origin is a no-op. - .agents/skills/outage-local-landing: the procedure firstmate follows on a health transition and before landing during an outage. Colocated tests cover the probe (confirmed-down, recovery, single-transient tolerance, boundedness, transition dedup), the watcher hook, the merge gates (refuse when up, accept during outage, diverged escalation, review gate, local-only back-compat), and sync (push+dispatch, idempotency, diverged refusal, undispatched-deferred note). * no-mistakes(document): docs: align outage-landing docs with the new scripts * no-mistakes(document): fix outage-landing ledger field list in skill doc * chore: re-trigger CI on the fork * chore: trigger CI now that fork workflows are enabled * fix(test): make outage-sync fixtures branch-default-independent The diverged-case fixture cloned the bare origin and pushed 'main', but on a runner whose init.defaultBranch is master the clone checked out master and the push failed with 'src refspec main does not match any'. Force the bare origin's HEAD to main and check out main explicitly in the clone so the fixtures work on any runner default. Reproduced and verified with init.defaultBranch=master. * no-mistakes(document): docs: list outage-landings ledger and github-down flag in AGENTS.md state inventory --------- Co-authored-by: Kun Chen <3233006+kunchenguid@users.noreply.github.com>
joliverMI
pushed a commit
to joliverMI/firstmate
that referenced
this pull request
Aug 17, 2026
…mlink (kunchenguid#2512) A Write aimed at CLAUDE.md followed the symlink and destroyed AGENTS.md. The installer now creates and migrates to a recoverable two-line pointer file.
joliverMI
added a commit
to joliverMI/firstmate
that referenced
this pull request
Aug 17, 2026
* feat(stow): add open-record persistence to /stow before reset (kunchenguid#2488) * feat(stow): persist the open records a session is holding /stow curated memory and captured session knowledge, but never touched record state, while AGENTS.md called it an "unfinished-work sweep" and the receipt declared the session "safe to reset" - wording that implied a record-correctness guarantee stow does not make. A shipped PR with no backlog item, a queued umbrella whose phases had merged, and four decision holds left open after their answers shipped all survived repeated stows. Add a bounded pass that files record state from the same volatile input the rest of stow already uses: the open threads in context, minutes before the reset destroys them. It creates a record for an unfiled thread and corrects one the session knows is wrong, through the owning path, and states its boundary as part of the contract - it never enumerates the backlog, lists holds, or queries a forge, because it cannot be a reconciliation and must not be read as one. Correct the wording in AGENTS.md and the completion receipt so reset-safe means what it actually guarantees: nothing this session knew was lost. * no-mistakes(review): correct stow decision-hold inspection to read hold via tasks-axi * no-mistakes(document): note /stow open-record persistence in README command catalog * refactor(stow): state open-record persistence as principle, not procedure The first version enumerated triggers, named commands, and prescribed an ordered procedure. That is too rigid for an agent skill: it invites literal execution of a checklist instead of judgment, and every enumerated example is a way for the guidance to go stale. Reduce it to the intent - before a reset, the important open work you are holding in context must end up durably recorded rather than dying with the session, filing what is unfiled and correcting what is stale - and let the agent judge importance, the record, and the owning write path. Keep the scope bound, since it is a decided contract and not a mechanic: this covers the open work the session is holding, never a reconciliation of durable records against repository or forge reality. The wording corrections in AGENTS.md and the completion receipt are unchanged. * fix(decisions): close decision holds at answer time via one general keyed-answer path (kunchenguid#2490) * fix(decisions): close captain holds at answer time Firstmate had two "a decision is open" ledgers with asymmetric closing mechanics. The live status-log ledger closes atomically at answer time, because bin/fm-send.sh --resolve-key makes answering a decision be the act that closes it. The durable backlog hold ledger had no such coupling: answering and recording were two separate acts, and only the first was forced by the workflow. That asymmetry lost four real captain decisions. Their answers were captured durably to disk, keyed character for character by the hold decision keys, acknowledged, and even implemented and shipped, yet the holds stayed open for two days and the captain was asked to re-answer decisions already on his own disk. Give the hold ledger the same answer-time-closure property: - bin/fm-decision-hold.sh gains an `answer` subcommand, the hold ledger's counterpart to --resolve-key. It shares one unrouted close implementation with `decline`, so it carries every existing guard - the captain decision file, the active-hold requirement, retry identity, and the refusal to release still-routed work - and differs only in the resolution mode it records. `decline` keeps its stronger meaning that the answer routes no follow-up work at all. - bin/fm-procevent-lavish.sh wires the channel that actually carried the lost answers. `arm --decisions-origin` binds a deck to the origin whose holds it carries, `answers` reads the structured choices out of a captured poll result, `close-decisions` maps each key to its hold and closes it through the command above, and `autohandle` lets the runner apply that at capture time. Safety is preserved rather than traded away. Only rows tagged `choice` are read, so freeform captain prose cannot forge a decision key. Closure is confined to the one bound origin. The decision text is a pure function of the captured result, so a replayed capture is idempotent. A hold that is absent, already closed, or still blocking routed work is skipped and left for `resolve`, never forced. A deck armed without the binding touches no hold at all. And autohandle deliberately never reports full handling, because recording an answer is transcription while acting on it is firstmate's judgement - so the check wake still reaches the handler. fm-send --resolve-key is untouched. * no-mistakes(document): document state/lavish-decisions binding dir in AGENTS.md state inventory * refactor(decisions): make keyed-answer closure one general capability The previous pass gave holds answer-time closure but built it as bespoke Lavish wiring: the review adapter carried the source-to-origin binding, mapped keys to hold identities, wrote decision records, decided what to skip, and closed holds itself. That treated a review deck as a special decision source. It is not - it is an ephemeral discussion format that happens to carry answers. Collapse it into ONE general capability with one owner. bin/fm-decision-hold.sh now owns the whole of "a keyed answer closes its matching hold": - `answers <origin> --source <provenance>` is the channel-agnostic intake. It reads key/answer/label lines on stdin, maps each key to its hold, and closes it through the same `answer` path, so every guard applies identically whatever channel the answer came from. --source is provenance recorded in the decision, never a behavior switch; there is no per-channel branch and no knowledge of chat, decks, or transports. - `bind`/`unbind`/`binding` own the source-to-origin binding for any channel whose answers arrive detached from their origin. Every channel is now an ordinary caller that only turns what it received into keyed lines: - bin/fm-send.sh (chat) feeds the intake for a key that names an active hold. This also fixes a real gap: once `complete` transfers a decision to its hold it closes the live status copy, so --resolve-key alone could never answer a transferred decision. - bin/fm-procevent.sh feeds it generically. A bound source's captured result goes to `<adapter> answers <result-file>` and whatever that prints is piped into the intake. The runner names no adapter, parses no result, and carries no decision rule, so any future adapter with an `answers` command works with no change here. - bin/fm-procevent-lavish.sh keeps only `answers`, which reports the structured choices a review captured and stops. It maps nothing to a hold and closes nothing; it lost ~160 lines of decision logic. Feeding is independent of handling, so it never acknowledges a result and never suppresses a wake - recording an answer is transcription, acting on it stays firstmate's judgement. The regression that proves closure now drives a FIXTURE adapter that is not the review adapter, so what is proven is that any bound channel reaches the intake rather than that one channel is wired specially. A new regression drives the real fm-send over a stubbed transport for the chat side. Every prior guarantee still holds, and fm-send's status-log behavior is unchanged. * no-mistakes(review): test(decisions): drop source-content grep from hold-closure regression * fix(memory): emit a real @AGENTS.md pointer instead of a CLAUDE.md symlink (kunchenguid#2512) A Write aimed at CLAUDE.md followed the symlink and destroyed AGENTS.md. The installer now creates and migrates to a recoverable two-line pointer file. * fix(ci): keep CLAUDE.md pointer check valid (kunchenguid#2515) * feat(dashboard): add the Admiral's Fleet Dashboard task board Replaces the generated Lavish status page with a purpose-built board: a zero-dependency Python backend (stdlib http.server + sqlite3), a vendored React frontend styled to match Spectra, an agent-facing CLI (bin/fm-dashboard.sh) that is the only way agents touch it, and the fleet-dashboard skill teaching the fleet auditor and other agents how to use it. The board owns its own persistent records rather than mirroring any backlog - its six-status vocabulary and four review tabs don't map onto tasks-axi state, and the fleet auditor needs a separate claim to check against live reality. See docs/dashboard.md for the full architecture, the deploy story, and the drift-risk tradeoff that choice implies. Link posting rejects GitHub/PR links and local-only hosts structurally (standing order 17), not just by agent memory. The page shows a loud, persistent banner when it cannot reach its own API, and the auditor's "never run" state is rendered distinctly from a clean sweep, so an absence of confirmation is never shown as a confirmed-good state. --------- Co-authored-by: Kun Chen <3233006+kunchenguid@users.noreply.github.com> Co-authored-by: joliverMI <joliver@sensibletech.biz>
wonder-media
added a commit
to wonder-media/firstmate
that referenced
this pull request
Aug 18, 2026
* fix: keep the public promise reachable when work is routed to a second mate (kunchenguid#2457) The lightweight Relay follow-up link lives in the answering home's own state/<task-id>.meta, so it can only bind work that home owns. When a Relay-linked request is routed to a second mate, the task record lives in the second mate's home, fm-x-link.sh failed with a bare "no such task ...meta", and nothing else picked the promise up: only the soft acknowledgement was ever posted. The typed promised-final path already supports --work-home secondmate:<id>; the playbook simply never chose it. - fmx-respond now states the routing rule crisply: a task in this home takes the lightweight link, and second-mate-routed work takes a promised-final commitment bound to that home, registered up front with the brief command carried into the routed worker's instructions. - fm-x-link.sh refuses a task with no local record by naming the registered second mate whose home actually holds it and printing the promised-final registration command, with the exact --work-home when the match is unambiguous. A home with no registered second mates keeps the plain error. - fm-backlog-handoff.sh reports, after a successful move, any moved key that still owes a public reply bound to main/<key>, since that binding no longer names the home owning the work. The move itself is never blocked. Docs and the secondmate handoff prose follow the same rule. Tests cover the refusal, its scoping, the unchanged local-link path, and both handoff outcomes at the script boundary. * docs(skills): add remote-secondmate recovery hint for false-negative verdicts (kunchenguid#2456) * fix(skills): hint that remote secondmate liveness verdicts false-negative fm-crew-state and fm-send routinely misreport a live remote secondmate as dead; confirm against the pane before relaunching, and relaunch only through fm-spawn.sh, never raw herdr pane surgery. * no-mistakes: apply CI fixes * fix(calm): keep Pi's export confirmation visible (kunchenguid#2461) Pi 0.83.0 added a status line to every tool-expansion change, and Pi updates the previous status line in place when two status messages arrive back to back. Calm's post-export redraw cycled tool expansion on the macrotask right after Pi printed "Session exported to: <path>", so both expansion status lines coalesced over that confirmation and the captain was left with no record of where their export landed. Calm now repaints only the tool rows it presents, by invalidating each row through the render context Pi hands its render slots, and requests the surrounding redraw through setStatus. Neither appends to the transcript. The repaint is still needed because Pi can re-render a row asynchronously - the built-in edit row invalidates itself once its diff is ready - and that re-render can land inside the window where /export forces stock rendering. The real-terminal /export case now asserts the confirmation is still on screen after the redraw has settled, and that the redraw restored every Calm-hidden row, instead of only racing the moment the confirmation first appeared. * feat(stow): add open-record persistence to /stow before reset (kunchenguid#2488) * feat(stow): persist the open records a session is holding /stow curated memory and captured session knowledge, but never touched record state, while AGENTS.md called it an "unfinished-work sweep" and the receipt declared the session "safe to reset" - wording that implied a record-correctness guarantee stow does not make. A shipped PR with no backlog item, a queued umbrella whose phases had merged, and four decision holds left open after their answers shipped all survived repeated stows. Add a bounded pass that files record state from the same volatile input the rest of stow already uses: the open threads in context, minutes before the reset destroys them. It creates a record for an unfiled thread and corrects one the session knows is wrong, through the owning path, and states its boundary as part of the contract - it never enumerates the backlog, lists holds, or queries a forge, because it cannot be a reconciliation and must not be read as one. Correct the wording in AGENTS.md and the completion receipt so reset-safe means what it actually guarantees: nothing this session knew was lost. * no-mistakes(review): correct stow decision-hold inspection to read hold via tasks-axi * no-mistakes(document): note /stow open-record persistence in README command catalog * refactor(stow): state open-record persistence as principle, not procedure The first version enumerated triggers, named commands, and prescribed an ordered procedure. That is too rigid for an agent skill: it invites literal execution of a checklist instead of judgment, and every enumerated example is a way for the guidance to go stale. Reduce it to the intent - before a reset, the important open work you are holding in context must end up durably recorded rather than dying with the session, filing what is unfiled and correcting what is stale - and let the agent judge importance, the record, and the owning write path. Keep the scope bound, since it is a decided contract and not a mechanic: this covers the open work the session is holding, never a reconciliation of durable records against repository or forge reality. The wording corrections in AGENTS.md and the completion receipt are unchanged. * fix(decisions): close decision holds at answer time via one general keyed-answer path (kunchenguid#2490) * fix(decisions): close captain holds at answer time Firstmate had two "a decision is open" ledgers with asymmetric closing mechanics. The live status-log ledger closes atomically at answer time, because bin/fm-send.sh --resolve-key makes answering a decision be the act that closes it. The durable backlog hold ledger had no such coupling: answering and recording were two separate acts, and only the first was forced by the workflow. That asymmetry lost four real captain decisions. Their answers were captured durably to disk, keyed character for character by the hold decision keys, acknowledged, and even implemented and shipped, yet the holds stayed open for two days and the captain was asked to re-answer decisions already on his own disk. Give the hold ledger the same answer-time-closure property: - bin/fm-decision-hold.sh gains an `answer` subcommand, the hold ledger's counterpart to --resolve-key. It shares one unrouted close implementation with `decline`, so it carries every existing guard - the captain decision file, the active-hold requirement, retry identity, and the refusal to release still-routed work - and differs only in the resolution mode it records. `decline` keeps its stronger meaning that the answer routes no follow-up work at all. - bin/fm-procevent-lavish.sh wires the channel that actually carried the lost answers. `arm --decisions-origin` binds a deck to the origin whose holds it carries, `answers` reads the structured choices out of a captured poll result, `close-decisions` maps each key to its hold and closes it through the command above, and `autohandle` lets the runner apply that at capture time. Safety is preserved rather than traded away. Only rows tagged `choice` are read, so freeform captain prose cannot forge a decision key. Closure is confined to the one bound origin. The decision text is a pure function of the captured result, so a replayed capture is idempotent. A hold that is absent, already closed, or still blocking routed work is skipped and left for `resolve`, never forced. A deck armed without the binding touches no hold at all. And autohandle deliberately never reports full handling, because recording an answer is transcription while acting on it is firstmate's judgement - so the check wake still reaches the handler. fm-send --resolve-key is untouched. * no-mistakes(document): document state/lavish-decisions binding dir in AGENTS.md state inventory * refactor(decisions): make keyed-answer closure one general capability The previous pass gave holds answer-time closure but built it as bespoke Lavish wiring: the review adapter carried the source-to-origin binding, mapped keys to hold identities, wrote decision records, decided what to skip, and closed holds itself. That treated a review deck as a special decision source. It is not - it is an ephemeral discussion format that happens to carry answers. Collapse it into ONE general capability with one owner. bin/fm-decision-hold.sh now owns the whole of "a keyed answer closes its matching hold": - `answers <origin> --source <provenance>` is the channel-agnostic intake. It reads key/answer/label lines on stdin, maps each key to its hold, and closes it through the same `answer` path, so every guard applies identically whatever channel the answer came from. --source is provenance recorded in the decision, never a behavior switch; there is no per-channel branch and no knowledge of chat, decks, or transports. - `bind`/`unbind`/`binding` own the source-to-origin binding for any channel whose answers arrive detached from their origin. Every channel is now an ordinary caller that only turns what it received into keyed lines: - bin/fm-send.sh (chat) feeds the intake for a key that names an active hold. This also fixes a real gap: once `complete` transfers a decision to its hold it closes the live status copy, so --resolve-key alone could never answer a transferred decision. - bin/fm-procevent.sh feeds it generically. A bound source's captured result goes to `<adapter> answers <result-file>` and whatever that prints is piped into the intake. The runner names no adapter, parses no result, and carries no decision rule, so any future adapter with an `answers` command works with no change here. - bin/fm-procevent-lavish.sh keeps only `answers`, which reports the structured choices a review captured and stops. It maps nothing to a hold and closes nothing; it lost ~160 lines of decision logic. Feeding is independent of handling, so it never acknowledges a result and never suppresses a wake - recording an answer is transcription, acting on it stays firstmate's judgement. The regression that proves closure now drives a FIXTURE adapter that is not the review adapter, so what is proven is that any bound channel reaches the intake rather than that one channel is wired specially. A new regression drives the real fm-send over a stubbed transport for the chat side. Every prior guarantee still holds, and fm-send's status-log behavior is unchanged. * no-mistakes(review): test(decisions): drop source-content grep from hold-closure regression * fix(memory): emit a real @AGENTS.md pointer instead of a CLAUDE.md symlink (kunchenguid#2512) A Write aimed at CLAUDE.md followed the symlink and destroyed AGENTS.md. The installer now creates and migrates to a recoverable two-line pointer file. * fix(ci): keep CLAUDE.md pointer check valid (kunchenguid#2515) * ci: gate GitHub workflows with pinned actionlint (kunchenguid#2517) * fix(lint): catch malformed GitHub workflows before merge A self-broken ci.yml cannot report its own breakage, so parse every workflow in the local lint path that no-mistakes already runs. * fix(lint): pin actionlint instead of Ruby for workflow lint A self-broken ci.yml still has to fail in the local lint path, and the named tool for that gate is actionlint, not a new Ruby runtime. * no-mistakes(document): Clarify pinned workflow lint documentation * fix: install pinned lint tools across supported platforms (kunchenguid#2546) * fix: install pinned shellcheck and actionlint on macOS and linux arm64 The installers were hardcoded to linux amd64 and sha256sum, so a Mac dev could not satisfy the refuse-on-mismatch lint gate. Select the official per-platform archive and checksum, and fall back to shasum -a 256. * no-mistakes(document): Document cross-platform pinned lint installers * docs: reconcile test-evidence docs with store_in_repo: true (kunchenguid#2548) .no-mistakes.yaml has set test.evidence.store_in_repo: true since kunchenguid#2355, but CONTRIBUTING.md, docs/configuration.md, and docs/architecture.md still described the old policy of keeping evidence out of the repo in a temp directory. The current no-mistakes behavior for store_in_repo: true is to publish each run's test evidence to the orphan no-mistakes/evidence branch and link it from the PR body. That branch shares no history with code branches, so evidence never enters a pushed feature branch or the default branch, and CI's tracked personal fleet paths rule stays accurate. Docs only. No change to .no-mistakes.yaml or any workflow. * docs: clarify test evidence branch storage (kunchenguid#2549) * docs: correct test evidence storage comment in .no-mistakes.yaml * no-mistakes: apply CI fixes * feat: add worker auto-compaction launch config * no-mistakes(document): docs: add crew-autocompact to AGENTS.md config file inventory --------- Co-authored-by: Kun Chen <3233006+kunchenguid@users.noreply.github.com>
wonder-media
added a commit
to wonder-media/firstmate
that referenced
this pull request
Aug 18, 2026
#2) * fix: keep the public promise reachable when work is routed to a second mate (kunchenguid#2457) The lightweight Relay follow-up link lives in the answering home's own state/<task-id>.meta, so it can only bind work that home owns. When a Relay-linked request is routed to a second mate, the task record lives in the second mate's home, fm-x-link.sh failed with a bare "no such task ...meta", and nothing else picked the promise up: only the soft acknowledgement was ever posted. The typed promised-final path already supports --work-home secondmate:<id>; the playbook simply never chose it. - fmx-respond now states the routing rule crisply: a task in this home takes the lightweight link, and second-mate-routed work takes a promised-final commitment bound to that home, registered up front with the brief command carried into the routed worker's instructions. - fm-x-link.sh refuses a task with no local record by naming the registered second mate whose home actually holds it and printing the promised-final registration command, with the exact --work-home when the match is unambiguous. A home with no registered second mates keeps the plain error. - fm-backlog-handoff.sh reports, after a successful move, any moved key that still owes a public reply bound to main/<key>, since that binding no longer names the home owning the work. The move itself is never blocked. Docs and the secondmate handoff prose follow the same rule. Tests cover the refusal, its scoping, the unchanged local-link path, and both handoff outcomes at the script boundary. * docs(skills): add remote-secondmate recovery hint for false-negative verdicts (kunchenguid#2456) * fix(skills): hint that remote secondmate liveness verdicts false-negative fm-crew-state and fm-send routinely misreport a live remote secondmate as dead; confirm against the pane before relaunching, and relaunch only through fm-spawn.sh, never raw herdr pane surgery. * no-mistakes: apply CI fixes * fix(calm): keep Pi's export confirmation visible (kunchenguid#2461) Pi 0.83.0 added a status line to every tool-expansion change, and Pi updates the previous status line in place when two status messages arrive back to back. Calm's post-export redraw cycled tool expansion on the macrotask right after Pi printed "Session exported to: <path>", so both expansion status lines coalesced over that confirmation and the captain was left with no record of where their export landed. Calm now repaints only the tool rows it presents, by invalidating each row through the render context Pi hands its render slots, and requests the surrounding redraw through setStatus. Neither appends to the transcript. The repaint is still needed because Pi can re-render a row asynchronously - the built-in edit row invalidates itself once its diff is ready - and that re-render can land inside the window where /export forces stock rendering. The real-terminal /export case now asserts the confirmation is still on screen after the redraw has settled, and that the redraw restored every Calm-hidden row, instead of only racing the moment the confirmation first appeared. * feat(stow): add open-record persistence to /stow before reset (kunchenguid#2488) * feat(stow): persist the open records a session is holding /stow curated memory and captured session knowledge, but never touched record state, while AGENTS.md called it an "unfinished-work sweep" and the receipt declared the session "safe to reset" - wording that implied a record-correctness guarantee stow does not make. A shipped PR with no backlog item, a queued umbrella whose phases had merged, and four decision holds left open after their answers shipped all survived repeated stows. Add a bounded pass that files record state from the same volatile input the rest of stow already uses: the open threads in context, minutes before the reset destroys them. It creates a record for an unfiled thread and corrects one the session knows is wrong, through the owning path, and states its boundary as part of the contract - it never enumerates the backlog, lists holds, or queries a forge, because it cannot be a reconciliation and must not be read as one. Correct the wording in AGENTS.md and the completion receipt so reset-safe means what it actually guarantees: nothing this session knew was lost. * no-mistakes(review): correct stow decision-hold inspection to read hold via tasks-axi * no-mistakes(document): note /stow open-record persistence in README command catalog * refactor(stow): state open-record persistence as principle, not procedure The first version enumerated triggers, named commands, and prescribed an ordered procedure. That is too rigid for an agent skill: it invites literal execution of a checklist instead of judgment, and every enumerated example is a way for the guidance to go stale. Reduce it to the intent - before a reset, the important open work you are holding in context must end up durably recorded rather than dying with the session, filing what is unfiled and correcting what is stale - and let the agent judge importance, the record, and the owning write path. Keep the scope bound, since it is a decided contract and not a mechanic: this covers the open work the session is holding, never a reconciliation of durable records against repository or forge reality. The wording corrections in AGENTS.md and the completion receipt are unchanged. * fix(decisions): close decision holds at answer time via one general keyed-answer path (kunchenguid#2490) * fix(decisions): close captain holds at answer time Firstmate had two "a decision is open" ledgers with asymmetric closing mechanics. The live status-log ledger closes atomically at answer time, because bin/fm-send.sh --resolve-key makes answering a decision be the act that closes it. The durable backlog hold ledger had no such coupling: answering and recording were two separate acts, and only the first was forced by the workflow. That asymmetry lost four real captain decisions. Their answers were captured durably to disk, keyed character for character by the hold decision keys, acknowledged, and even implemented and shipped, yet the holds stayed open for two days and the captain was asked to re-answer decisions already on his own disk. Give the hold ledger the same answer-time-closure property: - bin/fm-decision-hold.sh gains an `answer` subcommand, the hold ledger's counterpart to --resolve-key. It shares one unrouted close implementation with `decline`, so it carries every existing guard - the captain decision file, the active-hold requirement, retry identity, and the refusal to release still-routed work - and differs only in the resolution mode it records. `decline` keeps its stronger meaning that the answer routes no follow-up work at all. - bin/fm-procevent-lavish.sh wires the channel that actually carried the lost answers. `arm --decisions-origin` binds a deck to the origin whose holds it carries, `answers` reads the structured choices out of a captured poll result, `close-decisions` maps each key to its hold and closes it through the command above, and `autohandle` lets the runner apply that at capture time. Safety is preserved rather than traded away. Only rows tagged `choice` are read, so freeform captain prose cannot forge a decision key. Closure is confined to the one bound origin. The decision text is a pure function of the captured result, so a replayed capture is idempotent. A hold that is absent, already closed, or still blocking routed work is skipped and left for `resolve`, never forced. A deck armed without the binding touches no hold at all. And autohandle deliberately never reports full handling, because recording an answer is transcription while acting on it is firstmate's judgement - so the check wake still reaches the handler. fm-send --resolve-key is untouched. * no-mistakes(document): document state/lavish-decisions binding dir in AGENTS.md state inventory * refactor(decisions): make keyed-answer closure one general capability The previous pass gave holds answer-time closure but built it as bespoke Lavish wiring: the review adapter carried the source-to-origin binding, mapped keys to hold identities, wrote decision records, decided what to skip, and closed holds itself. That treated a review deck as a special decision source. It is not - it is an ephemeral discussion format that happens to carry answers. Collapse it into ONE general capability with one owner. bin/fm-decision-hold.sh now owns the whole of "a keyed answer closes its matching hold": - `answers <origin> --source <provenance>` is the channel-agnostic intake. It reads key/answer/label lines on stdin, maps each key to its hold, and closes it through the same `answer` path, so every guard applies identically whatever channel the answer came from. --source is provenance recorded in the decision, never a behavior switch; there is no per-channel branch and no knowledge of chat, decks, or transports. - `bind`/`unbind`/`binding` own the source-to-origin binding for any channel whose answers arrive detached from their origin. Every channel is now an ordinary caller that only turns what it received into keyed lines: - bin/fm-send.sh (chat) feeds the intake for a key that names an active hold. This also fixes a real gap: once `complete` transfers a decision to its hold it closes the live status copy, so --resolve-key alone could never answer a transferred decision. - bin/fm-procevent.sh feeds it generically. A bound source's captured result goes to `<adapter> answers <result-file>` and whatever that prints is piped into the intake. The runner names no adapter, parses no result, and carries no decision rule, so any future adapter with an `answers` command works with no change here. - bin/fm-procevent-lavish.sh keeps only `answers`, which reports the structured choices a review captured and stops. It maps nothing to a hold and closes nothing; it lost ~160 lines of decision logic. Feeding is independent of handling, so it never acknowledges a result and never suppresses a wake - recording an answer is transcription, acting on it stays firstmate's judgement. The regression that proves closure now drives a FIXTURE adapter that is not the review adapter, so what is proven is that any bound channel reaches the intake rather than that one channel is wired specially. A new regression drives the real fm-send over a stubbed transport for the chat side. Every prior guarantee still holds, and fm-send's status-log behavior is unchanged. * no-mistakes(review): test(decisions): drop source-content grep from hold-closure regression * fix(memory): emit a real @AGENTS.md pointer instead of a CLAUDE.md symlink (kunchenguid#2512) A Write aimed at CLAUDE.md followed the symlink and destroyed AGENTS.md. The installer now creates and migrates to a recoverable two-line pointer file. * fix: tighten generated brief evidence discipline --------- Co-authored-by: Kun Chen <3233006+kunchenguid@users.noreply.github.com>
digbycampbell
added a commit
to digbycampbell/firstmate
that referenced
this pull request
Aug 18, 2026
… decision-hold bindings (#8) * fix: keep the public promise reachable when work is routed to a second mate (kunchenguid#2457) The lightweight Relay follow-up link lives in the answering home's own state/<task-id>.meta, so it can only bind work that home owns. When a Relay-linked request is routed to a second mate, the task record lives in the second mate's home, fm-x-link.sh failed with a bare "no such task ...meta", and nothing else picked the promise up: only the soft acknowledgement was ever posted. The typed promised-final path already supports --work-home secondmate:<id>; the playbook simply never chose it. - fmx-respond now states the routing rule crisply: a task in this home takes the lightweight link, and second-mate-routed work takes a promised-final commitment bound to that home, registered up front with the brief command carried into the routed worker's instructions. - fm-x-link.sh refuses a task with no local record by naming the registered second mate whose home actually holds it and printing the promised-final registration command, with the exact --work-home when the match is unambiguous. A home with no registered second mates keeps the plain error. - fm-backlog-handoff.sh reports, after a successful move, any moved key that still owes a public reply bound to main/<key>, since that binding no longer names the home owning the work. The move itself is never blocked. Docs and the secondmate handoff prose follow the same rule. Tests cover the refusal, its scoping, the unchanged local-link path, and both handoff outcomes at the script boundary. * docs(skills): add remote-secondmate recovery hint for false-negative verdicts (kunchenguid#2456) * fix(skills): hint that remote secondmate liveness verdicts false-negative fm-crew-state and fm-send routinely misreport a live remote secondmate as dead; confirm against the pane before relaunching, and relaunch only through fm-spawn.sh, never raw herdr pane surgery. * no-mistakes: apply CI fixes * fix(calm): keep Pi's export confirmation visible (kunchenguid#2461) Pi 0.83.0 added a status line to every tool-expansion change, and Pi updates the previous status line in place when two status messages arrive back to back. Calm's post-export redraw cycled tool expansion on the macrotask right after Pi printed "Session exported to: <path>", so both expansion status lines coalesced over that confirmation and the captain was left with no record of where their export landed. Calm now repaints only the tool rows it presents, by invalidating each row through the render context Pi hands its render slots, and requests the surrounding redraw through setStatus. Neither appends to the transcript. The repaint is still needed because Pi can re-render a row asynchronously - the built-in edit row invalidates itself once its diff is ready - and that re-render can land inside the window where /export forces stock rendering. The real-terminal /export case now asserts the confirmation is still on screen after the redraw has settled, and that the redraw restored every Calm-hidden row, instead of only racing the moment the confirmation first appeared. * feat(stow): add open-record persistence to /stow before reset (kunchenguid#2488) * feat(stow): persist the open records a session is holding /stow curated memory and captured session knowledge, but never touched record state, while AGENTS.md called it an "unfinished-work sweep" and the receipt declared the session "safe to reset" - wording that implied a record-correctness guarantee stow does not make. A shipped PR with no backlog item, a queued umbrella whose phases had merged, and four decision holds left open after their answers shipped all survived repeated stows. Add a bounded pass that files record state from the same volatile input the rest of stow already uses: the open threads in context, minutes before the reset destroys them. It creates a record for an unfiled thread and corrects one the session knows is wrong, through the owning path, and states its boundary as part of the contract - it never enumerates the backlog, lists holds, or queries a forge, because it cannot be a reconciliation and must not be read as one. Correct the wording in AGENTS.md and the completion receipt so reset-safe means what it actually guarantees: nothing this session knew was lost. * no-mistakes(review): correct stow decision-hold inspection to read hold via tasks-axi * no-mistakes(document): note /stow open-record persistence in README command catalog * refactor(stow): state open-record persistence as principle, not procedure The first version enumerated triggers, named commands, and prescribed an ordered procedure. That is too rigid for an agent skill: it invites literal execution of a checklist instead of judgment, and every enumerated example is a way for the guidance to go stale. Reduce it to the intent - before a reset, the important open work you are holding in context must end up durably recorded rather than dying with the session, filing what is unfiled and correcting what is stale - and let the agent judge importance, the record, and the owning write path. Keep the scope bound, since it is a decided contract and not a mechanic: this covers the open work the session is holding, never a reconciliation of durable records against repository or forge reality. The wording corrections in AGENTS.md and the completion receipt are unchanged. * fix(decisions): close decision holds at answer time via one general keyed-answer path (kunchenguid#2490) * fix(decisions): close captain holds at answer time Firstmate had two "a decision is open" ledgers with asymmetric closing mechanics. The live status-log ledger closes atomically at answer time, because bin/fm-send.sh --resolve-key makes answering a decision be the act that closes it. The durable backlog hold ledger had no such coupling: answering and recording were two separate acts, and only the first was forced by the workflow. That asymmetry lost four real captain decisions. Their answers were captured durably to disk, keyed character for character by the hold decision keys, acknowledged, and even implemented and shipped, yet the holds stayed open for two days and the captain was asked to re-answer decisions already on his own disk. Give the hold ledger the same answer-time-closure property: - bin/fm-decision-hold.sh gains an `answer` subcommand, the hold ledger's counterpart to --resolve-key. It shares one unrouted close implementation with `decline`, so it carries every existing guard - the captain decision file, the active-hold requirement, retry identity, and the refusal to release still-routed work - and differs only in the resolution mode it records. `decline` keeps its stronger meaning that the answer routes no follow-up work at all. - bin/fm-procevent-lavish.sh wires the channel that actually carried the lost answers. `arm --decisions-origin` binds a deck to the origin whose holds it carries, `answers` reads the structured choices out of a captured poll result, `close-decisions` maps each key to its hold and closes it through the command above, and `autohandle` lets the runner apply that at capture time. Safety is preserved rather than traded away. Only rows tagged `choice` are read, so freeform captain prose cannot forge a decision key. Closure is confined to the one bound origin. The decision text is a pure function of the captured result, so a replayed capture is idempotent. A hold that is absent, already closed, or still blocking routed work is skipped and left for `resolve`, never forced. A deck armed without the binding touches no hold at all. And autohandle deliberately never reports full handling, because recording an answer is transcription while acting on it is firstmate's judgement - so the check wake still reaches the handler. fm-send --resolve-key is untouched. * no-mistakes(document): document state/lavish-decisions binding dir in AGENTS.md state inventory * refactor(decisions): make keyed-answer closure one general capability The previous pass gave holds answer-time closure but built it as bespoke Lavish wiring: the review adapter carried the source-to-origin binding, mapped keys to hold identities, wrote decision records, decided what to skip, and closed holds itself. That treated a review deck as a special decision source. It is not - it is an ephemeral discussion format that happens to carry answers. Collapse it into ONE general capability with one owner. bin/fm-decision-hold.sh now owns the whole of "a keyed answer closes its matching hold": - `answers <origin> --source <provenance>` is the channel-agnostic intake. It reads key/answer/label lines on stdin, maps each key to its hold, and closes it through the same `answer` path, so every guard applies identically whatever channel the answer came from. --source is provenance recorded in the decision, never a behavior switch; there is no per-channel branch and no knowledge of chat, decks, or transports. - `bind`/`unbind`/`binding` own the source-to-origin binding for any channel whose answers arrive detached from their origin. Every channel is now an ordinary caller that only turns what it received into keyed lines: - bin/fm-send.sh (chat) feeds the intake for a key that names an active hold. This also fixes a real gap: once `complete` transfers a decision to its hold it closes the live status copy, so --resolve-key alone could never answer a transferred decision. - bin/fm-procevent.sh feeds it generically. A bound source's captured result goes to `<adapter> answers <result-file>` and whatever that prints is piped into the intake. The runner names no adapter, parses no result, and carries no decision rule, so any future adapter with an `answers` command works with no change here. - bin/fm-procevent-lavish.sh keeps only `answers`, which reports the structured choices a review captured and stops. It maps nothing to a hold and closes nothing; it lost ~160 lines of decision logic. Feeding is independent of handling, so it never acknowledges a result and never suppresses a wake - recording an answer is transcription, acting on it stays firstmate's judgement. The regression that proves closure now drives a FIXTURE adapter that is not the review adapter, so what is proven is that any bound channel reaches the intake rather than that one channel is wired specially. A new regression drives the real fm-send over a stubbed transport for the chat side. Every prior guarantee still holds, and fm-send's status-log behavior is unchanged. * no-mistakes(review): test(decisions): drop source-content grep from hold-closure regression * fix(memory): emit a real @AGENTS.md pointer instead of a CLAUDE.md symlink (kunchenguid#2512) A Write aimed at CLAUDE.md followed the symlink and destroyed AGENTS.md. The installer now creates and migrates to a recoverable two-line pointer file. * fix(ci): keep CLAUDE.md pointer check valid (kunchenguid#2515) * ci: gate GitHub workflows with pinned actionlint (kunchenguid#2517) * fix(lint): catch malformed GitHub workflows before merge A self-broken ci.yml cannot report its own breakage, so parse every workflow in the local lint path that no-mistakes already runs. * fix(lint): pin actionlint instead of Ruby for workflow lint A self-broken ci.yml still has to fail in the local lint path, and the named tool for that gate is actionlint, not a new Ruby runtime. * no-mistakes(document): Clarify pinned workflow lint documentation * fix: install pinned lint tools across supported platforms (kunchenguid#2546) * fix: install pinned shellcheck and actionlint on macOS and linux arm64 The installers were hardcoded to linux amd64 and sha256sum, so a Mac dev could not satisfy the refuse-on-mismatch lint gate. Select the official per-platform archive and checksum, and fall back to shasum -a 256. * no-mistakes(document): Document cross-platform pinned lint installers * docs: reconcile test-evidence docs with store_in_repo: true (kunchenguid#2548) .no-mistakes.yaml has set test.evidence.store_in_repo: true since kunchenguid#2355, but CONTRIBUTING.md, docs/configuration.md, and docs/architecture.md still described the old policy of keeping evidence out of the repo in a temp directory. The current no-mistakes behavior for store_in_repo: true is to publish each run's test evidence to the orphan no-mistakes/evidence branch and link it from the PR body. That branch shares no history with code branches, so evidence never enters a pushed feature branch or the default branch, and CI's tracked personal fleet paths rule stays accurate. Docs only. No change to .no-mistakes.yaml or any workflow. * docs: clarify test evidence branch storage (kunchenguid#2549) * docs: correct test evidence storage comment in .no-mistakes.yaml * no-mistakes: apply CI fixes * fix(stow): take upstream's description wording verbatim in the merge resolution The conflict resolution had kept our dropped 'e.g.'; the captain's standing direction is to prefer upstream wording where the two genuinely conflict. Only the block-scalar reflow remains ours, so the file keeps one sentence per line as the repo style requires. --------- Co-authored-by: Kun Chen <3233006+kunchenguid@users.noreply.github.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Intent
Root-cause fix: make the agent-memory installer emit a real @AGENTS.md pointer, not a CLAUDE.md symlink.
Captain-approved (2026-08-16). Context: an agent Write aimed at CLAUDE.md followed the CLAUDE.md -> AGENTS.md symlink and destroyed 1488 lines of AGENTS.md. The fleet-wide decision is to replace that symlink convention with a real CLAUDE.md file that uses Claude Code's native @path import, which inlines AGENTS.md into memory at load time (zero extra read turns) while a stray write only clobbers a recoverable one-liner. A separate worker is swapping the 34 existing repos; THIS work is the root cause so the footgun never regenerates.
The canonical CLAUDE.md pointer content is the source of truth for this task and must be byte-identical to this exact two-line form (kept consistent with the fleet sweep):
@AGENTS.md
Required changes:
Acceptance criteria:
Do not merge the PR. Report the PR URL when CI is green.
What Changed
bin/fm-ensure-agents-md.shso it never creates aCLAUDE.mdsymlink: it now writes/asserts a canonical two-line real@AGENTS.mdpointer file, migrates an existing correctCLAUDE.md -> AGENTS.mdsymlink to that pointer (unlinking first so a write can't follow the link and destroyAGENTS.md), accepts the pointer as correct in the "both real files" branch, treats a second run as a clean no-op, and preserves every existing conflict/case-variant refusal (header, help, and echo strings updated accordingly).CLAUDE.mdsymlink with the canonical real pointer file and corrected the symlink-convention wording inAGENTS.md,CONTRIBUTING.md,docs/architecture.md,docs/scripts.md,bin/fm-ff-lib.sh, and theupdatefirstmateskill; the CI "symlinks intact" check now enforces thatCLAUDE.mdis a real file byte-equal to the canonical pointer rather than a symlink.tests/fm-ensure-agents-md.test.shto exercise the real script for pointer creation on fresh setup, symlink→pointer migration without clobberingAGENTS.md, no-op idempotency, and the preserved conflict refusals.Risk Assessment
✅ Low: The change fully satisfies every required intent constraint with byte-identical canonical content, an unlink-before-write path that closes the symlink footgun, all safety refusals preserved, consistent doc/CI updates, and behavior-exercising tests; every branch traces correctly with no reachable wrong-result path.
Testing
Ran the full focused test suite for the installer (16/16 passing, all exercising the real executable), then demonstrated the core intent end-to-end by contrasting the base-commit symlink installer (a stray write to CLAUDE.md destroyed AGENTS.md) against the target-commit pointer installer (the identical stray write left AGENTS.md intact), plus symlink->pointer migration, clean no-op idempotency, and byte-identical canonical pointer content; also confirmed the repo's own CLAUDE.md is now a real pointer file (git-recorded symlink->file mode change) and no lingering symlink-convention docs remain. This is a CLI/filesystem-behavior change with no rendered UI surface, so evidence is CLI transcripts and file-state inspection rather than screenshots. Overall result: green.Evidence: E2E footgun comparison (old symlink destroys AGENTS.md vs new pointer protects it)
Source: E2E footgun comparison (old symlink destroys AGENTS.md vs new pointer protects it)
Evidence: E2E symlink->pointer migration and idempotency transcript
Source: E2E symlink->pointer migration and idempotency transcript
Pipeline
Updates from git push no-mistakes
✅ **intent** - passed
✅ No issues found.
✅ **Rebase** - passed
✅ No issues found.
✅ **Review** - passed
✅ No issues found.
✅ **Test** - passed
✅ No issues found.
bash tests/fm-ensure-agents-md.test.sh(16/16 ok)E2E footgun comparison: ran base-commit vs target-commit installer, thenecho 'stray note' > CLAUDE.mdin each and inspected AGENTS.md survivalE2E migration: correct CLAUDE.md->AGENTS.md symlink migrated to real pointer with AGENTS.md payload preserved, second run reportsunchanged:E2E idempotency: three consecutivebin/fm-ensure-agents-md.shruns on a fresh repo (created -> unchanged -> unchanged)Verified repo CLAUDE.md is a regular file byte-identical to the canonical two-line pointer viacmpgit diffconfirms repo CLAUDE.md changed from file mode 120000 (symlink) to 100644 (regular file)grep scan of docs/AGENTS.md/CONTRIBUTING.md/README.md/.agents for lingering CLAUDE.md-symlink descriptions (none)✅ **Document** - passed
✅ No issues found.
✅ **Lint** - passed
✅ No issues found.
✅ **Push** - passed
✅ No issues found.