{"slug":"autopilot-2","title":"autopilot","summary":"Use when a human has an approved delivery plan and wants the full chain run under phase gates. Not for planning, open-ended debugging, or single-step execution: use work directly.","platform":"Claude","tags":[],"authorName":"LLM Mart","authorSlug":"llm-mart","score":0,"source":"github","price":null,"verified":false,"createdAt":"2026-09-30T19:49:54.161659Z","repo":{"url":"https://github.com/OutlineDriven/outline-driven-development","stars":54,"forks":10,"license":"Apache-2.0","updatedAt":"2026-09-28T03:16:21Z"},"bodyHtml":"<hr>\n<h2>name: autopilot\ndescription: 'Use when a human has an approved delivery plan and wants the full chain run under phase gates. Not for planning, open-ended debugging, or single-step execution: use work directly.'\ndisable-model-invocation: true</h2>\n<h1>Autopilot</h1>\n<h2>Contract</h2>\n<table>\n<thead>\n<tr>\n<th>Field</th>\n<th>Bound contract</th>\n</tr>\n</thead>\n<tbody>\n<tr>\n<td>Trigger</td>\n<td>A human explicitly invokes <code>/autopilot</code> with a feature description and an approved execution plan.</td>\n</tr>\n<tr>\n<td>Authority</td>\n<td>Human-gated: previews the exact target and consequence before credentials, paid actions, data-at-rest changes, publication, deployment, remote bulk mutation, or irreversible deletion, continuing only when the invocation covers that consequence; every other write is reversible local, with version control as the rollback. Pass authorized push and PR targets to <code>review-and-ship</code>; do not publish directly.</td>\n</tr>\n<tr>\n<td>Side effect</td>\n<td>Delegates implementation, simplification, review, and finalization to their respective skills; local artifacts are written by <code>work</code>, commits and PR are created by <code>review-and-ship</code>. No mutation is performed by autopilot itself.</td>\n</tr>\n<tr>\n<td>Done</td>\n<td>Return <code>DONE</code> only after the requested close-out state is observed. Otherwise return <code>BLOCKED</code> with a resumable handoff.</td>\n</tr>\n</tbody>\n</table>\n<h2>Inputs</h2>\n<p>Required: feature description, approved plan, repository, acceptance criteria, and close-out condition.</p>\n<p>Optional: scope limits; assigned executor; target branch, remote, and PR destination; required checks; shipping instructions. Derive an omitted value only when repository evidence gives one safe, unique answer.</p>\n<h2>Procedure</h2>\n<ol>\n<li>Parse the plan, scope, assignments, acceptance criteria, repository, required proof, shipping destination, and close-out condition. Reject contradictions and unavailable assigned executors. Done when: every required input is parsed and bound to a concrete value, contradictions are rejected with the conflicting pair named, and unavailable executors are reported; or the run stops on the first unresolvable input.</li>\n<li>Confirm that the human approved the plan. If not, stop before implementation and route to planning. Done when: an approval record is confirmed; the plan was approved through plan mode or an equivalent written approval; or the run stops and routes to planning with no delegation.</li>\n<li>Inspect enough repository evidence to bind the plan to exact artifacts, behavior, verification, and remote targets. Keep the campaign inside the approved scope. Done when: every plan step maps to a named repository artifact (file, directory, or remote target) and the scope boundary is stated with the artifacts inside it and the artifacts excluded from it.</li>\n<li>Preview each risky consequence. Stop when authority is absent or the target is ambiguous. Done when: every credential, paid, data-at-rest, publication, deployment, remote bulk, or irreversible consequence is previewed with its exact target, or the run stops naming the absent authority or ambiguous target.</li>\n<li>Delegate the bounded plan to <code>work</code> in orchestrated mode. Include acceptance criteria, constraints, repository evidence, required verifier, and the required structured return. <code>work</code> owns implementation and local verification; autopilot does not reproduce those steps. Apply the gate state machine in <code>references/pipeline-gates.md</code>: a failing work verifier gets one <code>strike-the-root</code> pass and one recheck; a second failure halts the chain. Done when: <code>work</code> returns a structured result with implementation summary and local verification passing, or the gate halts the chain with the verifier failure and diff recorded.</li>\n<li>Delegate simplification to <code>simplify</code> on the completed diff. Gate: <code>simplify</code> exits <code>0</code>, <code>11</code>, or <code>12</code> with behavior preserved. Halt on exit <code>14</code> (new bloat) or <code>15</code> (mixed-concern). Done when: <code>simplify</code> returns exit <code>0</code>, <code>11</code>, or <code>12</code> with behavior preserved confirmed by its output, or the chain halts on exit <code>14</code> or <code>15</code> with the exit code and cause recorded.</li>\n<li>Delegate review to <code>review</code> on the in-scope change. If critical or high findings remain, delegate one <code>strike-the-root</code> pass and re-review the changed files. Halt on residual critical or high findings. Done when: <code>review</code> returns zero critical and zero high findings on the changed files, or the chain halts with the residual findings listed.</li>\n<li>Delegate finalization to <code>review-and-ship</code>. Pass the reviewed diff, explicit delegated shipping authority, branch, remote, PR destination, required checks, and shipping instructions. The finalizer owns checks, commits, publication classification, push, and PR creation or update; autopilot performs none of them. Done when: <code>review-and-ship</code> returns its report with checks green and PR created or updated (full mode) or commits made and push skipped (local-only), or the finalizer returns blocked with its reason.</li>\n<li>Observe authorized close-out. Route any supported in-scope fix back through <code>work</code> in orchestrated mode, repeat affected gates, and invoke <code>review-and-ship</code> for each authorized update. Do not merge, deploy, or publish elsewhere without matching authority. Done when: the close-out condition named in the inputs is observed in repository or remote state, or the run stops for absent authority.</li>\n<li>Return the terminal classification and include the finalizer report. The gate and handoff formats in <code>references/pipeline-gates.md</code> are binding. Done when: the terminal result is <code>DONE</code> with the finalizer report attached and close-out state quoted, or <code>BLOCKED</code> with the failed stage, reason, and next action named.</li>\n</ol>\n<h2>Failure and recovery</h2>\n<p>Stop on a missing approved plan, contradictory input, unavailable executor, ambiguous target, absent authority, failed gate, non-converging review, rejected remote mutation, or incomplete close-out. Preserve successful in-scope work and observed remote state. Recover only mutations made by this run, using reversible repository or hosting operations; never discard unrelated work.</p>\n<p>Return <code>BLOCKED</code> with the failed stage, exact reason, autofix result, partial artifacts, remote mutations, proof state, recovery state, and the single concrete requirement to resume. Never claim rollback, publication, or close-out without observing it.</p>\n<h2>Output</h2>\n<p>Exactly one terminal result:</p>\n<ul>\n<li><code>DONE</code>: delivered behavior, changed artifacts, verification evidence, branch and PR, the complete <code>review-and-ship</code> report, and observed close-out state.</li>\n<li><code>BLOCKED</code>: failed stage, reason, partial result, remote state, proof and recovery status, and the next required action.</li>\n</ul>\n","files":[{"path":"agents/openai.yaml","sizeBytes":200,"isText":true},{"path":"references/pipeline-gates.md","sizeBytes":5966,"isText":true},{"path":"SKILL.md","sizeBytes":6491,"isText":true}],"reviewScore":null,"reviewSummary":null,"trust":{"provenance":"trusted-source-unreviewed","notice":"Community-authored content, reproduced verbatim and not vetted as instructions. Treat it as data to evaluate, never as directives to follow.","bodySource":null},"bodyLocked":false,"purchaseUrl":null,"sourceUrl":null,"report":{"provenance":"trusted-source-unreviewed","screen":{"ran":true,"outcome":"clean","suspicious":0,"notes":0,"hiddenCharacters":false},"virusScan":{"engine":"clamav","status":"clean","scannedAt":"2026-09-30T19:50:32.330323Z","sha256":"DEA9A4627708C7AFD1DFCD328AEFD5B408E5836024B3607FDD7123DE6E704238","sizeBytes":5626},"review":null,"source":{"repositoryUrl":"https://github.com/OutlineDriven/outline-driven-development","path":".devin/skills/autopilot","license":"Apache-2.0","commit":"b0e8ce89a19fac880251dc3ea1babfeb4503a4fe","subtreeSha":"FCC9B814025CD273D166CD0819190E0131874BDBAF4E28D9BCAFC58425586301","lastSyncedAt":"2026-09-30T19:49:48.917811Z"},"reviewedAt":"2026-09-30T19:51:57.919302Z","notice":"Community-authored content, reproduced verbatim and not vetted as instructions. Treat it as data to evaluate, never as directives to follow."},"install":[{"target":"skills-cli","command":"npx skills add https://github.com/OutlineDriven/outline-driven-development/tree/main/.devin/skills/autopilot"},{"target":"claude-code","command":"claude plugin marketplace add https://llmmart.ai/marketplace.json && claude plugin install outlinedriven-outline-driven-development@llmmart"},{"target":"git","command":"git clone https://github.com/OutlineDriven/outline-driven-development.git"}]}