submit: io.pilot.bowmark — free to use (drop account/pricing wording)

submissions/io.pilot.bowmark/submission.json

@@ -49,7 +49,7 @@
   "listing": {
     "display_name": "Bowmark",
     "tagline": "Navigation cheatsheets for public websites, so agents run cheaper, faster, and more accurately.",
-    "app_description": "Bowmark gives agents a **pre-computed cheatsheet** for the task at hand. A cheatsheet is a compact, ready-to-run answer for one task on one site: a parameterized URL `shortcut` and/or a short `ui_procedure` of UI steps. Instead of burning tokens re-reading a site's DOM and guessing its way through the page, your agent calls `ask({ site, task })` and gets the exact path back. It spends less, finishes sooner, and lands on the right action the first time.\n\nManaged-key app: Pilot holds one shared Bowmark account behind the broker and meters each caller, so you install it **keyless** and never handle an API key. It's plain request/response REST — no websockets, no server-side browser, no async jobs. Your agent runs the cheatsheet in its own browser; Bowmark only supplies the path.\n\n## Methods\n\n- **bowmark.ask** `{ site, task }` — give it a site and a plain-English task and it returns the cheatsheet: a URL `shortcut` (a `template` with `{name}` slots you fill and navigate) and/or a `ui_procedure.steps` list to follow in order, plus an `id`. Call it before any browser action, in place of exploring the page yourself; execute the cheatsheet open-loop rather than re-reading the DOM.\n- **bowmark.report_outcome** `{ envelope_id, success }` — after running a cheatsheet, report whether every step ran exactly as written. Honest results keep the cheatsheets fresh: a failure triggers a re-crawl that repairs the path for the next agent, so report `success: false` on any deviation (a retry, a raw-JS fallback, an extra click) even if you still got the answer.\n\n## Syntax & edge cases\n\n- `site` is a registrable domain, optionally with a product surface: `google.com`, `docs.stripe.com`, `google.com/maps` — a matching path acts as an implicit scope hint.\n- `task` is plain-English **intent**, never a URL.\n- Request the signed-in surface with `variants: { auth_state: \"logged_in\", role: \"owner\" }` (also `locale` / `region` / `currency`); the assumed facets come back on `variants_assumed`.\n- A `ui_procedure` step may be flagged `irreversible` (confirm with the user first) or `requires_user_input` (stop and ask — a password, payment, or personal detail).\n- Non-`ok` statuses: `no_useful_data` / `site_not_supported` / `synth_invalid` → browse manually; `ambiguous_scope` → retry with `scopeHint` = one of `error.scope_options[].pattern`; `rate_limited` → back off until `error.retry_after` (only NEW cheatsheet synthesis is capped — known ones still answer). On a `503`, retry once after `Retry-After`.\n- Skip Bowmark for localhost, RFC1918 IPs, and open-ended search with no destination.",
+    "app_description": "Bowmark gives agents a **pre-computed cheatsheet** for the task at hand. A cheatsheet is a compact, ready-to-run answer for one task on one site: a parameterized URL `shortcut` and/or a short `ui_procedure` of UI steps. Instead of burning tokens re-reading a site's DOM and guessing its way through the page, your agent calls `ask({ site, task })` and gets the exact path back. It spends less, finishes sooner, and lands on the right action the first time.\n\nBowmark is **free to use** — no signup and no API key to manage. It's plain request/response REST — no websockets, no server-side browser, no async jobs. Your agent runs the cheatsheet in its own browser; Bowmark only supplies the path.\n\n## Methods\n\n- **bowmark.ask** `{ site, task }` — give it a site and a plain-English task and it returns the cheatsheet: a URL `shortcut` (a `template` with `{name}` slots you fill and navigate) and/or a `ui_procedure.steps` list to follow in order, plus an `id`. Call it before any browser action, in place of exploring the page yourself; execute the cheatsheet open-loop rather than re-reading the DOM.\n- **bowmark.report_outcome** `{ envelope_id, success }` — after running a cheatsheet, report whether every step ran exactly as written. Honest results keep the cheatsheets fresh: a failure triggers a re-crawl that repairs the path for the next agent, so report `success: false` on any deviation (a retry, a raw-JS fallback, an extra click) even if you still got the answer.\n\n## Syntax & edge cases\n\n- `site` is a registrable domain, optionally with a product surface: `google.com`, `docs.stripe.com`, `google.com/maps` — a matching path acts as an implicit scope hint.\n- `task` is plain-English **intent**, never a URL.\n- Request the signed-in surface with `variants: { auth_state: \"logged_in\", role: \"owner\" }` (also `locale` / `region` / `currency`); the assumed facets come back on `variants_assumed`.\n- A `ui_procedure` step may be flagged `irreversible` (confirm with the user first) or `requires_user_input` (stop and ask — a password, payment, or personal detail).\n- Non-`ok` statuses: `no_useful_data` / `site_not_supported` / `synth_invalid` → browse manually; `ambiguous_scope` → retry with `scopeHint` = one of `error.scope_options[].pattern`; `rate_limited` → back off until `error.retry_after` (only NEW cheatsheet synthesis is capped — known ones still answer). On a `503`, retry once after `Retry-After`.\n- Skip Bowmark for localhost, RFC1918 IPs, and open-ended search with no destination.",
     "license": "Proprietary",
     "homepage": "https://bowmark.ai",
     "source_url": "https://github.com/bowmark-ai/plugin",