karloows/orchraft

Have the AI draft or update a standalone Markdown doc — a feature write-up, design doc, or a section of README/ROADMAP — grounded in the real diff, code, commit history, or PR/issue text. Use when the user says chronicle, asks to write documentation, draft a design doc, or update a named doc file such as README or ROADMAP.

Have the AI draft release notes for a tag or the unreleased work since the last one, grounded in the pull requests and commits that actually landed, then write them into the GitHub release once the user approves. Use when the user says herald, asks for release notes, or wants a release described.

Have the AI finish a topic branch end to end by merging its PR, removing the branch, and syncing the local base branch. Use when the user says land or asks to run the repo landing workflow.

Have the AI add or fix code comments and docstrings across a diff or file so they match this repo's lore policy, in whatever language or framework it's written in. Use when the user says lore, asks to comment the code, add docstrings, document a function, or clean up stale/noisy comments.

Have the AI report this repo's real shipping signal — PRs opened, review findings roast caught and whether they were fixed or declined, and time-to-land — grounded in live git/GitHub history, not an invented score. Use when the user says plunder, asks for shipping stats, wants a retrospective on PRs or reviews, or wants to know how the clan's actually been doing.

Have the AI triage, create, or update a GitHub issue before implementation starts. Use when the user says quest, asks to file a bug, create an issue, triage a report, or update an issue's status, labels, or body.

Have the AI list every review finding that was consciously declined rather than fixed, across the whole repo's pull requests, grounded in live GitHub data. Use when the user says reckoning, asks what's been declined, wants a debt list, or wants to see deferred findings before a release.

Have the AI review the current pull request's diff against the target repository's own policies and commit history — not an external ruleset — then post the findings as a PR review with inline comments and a summary. Use when the user says roast, review, asks for a PR review, or wants CodeRabbit-style automated review after shipping.

Have the AI summarize where the current branch's work actually stands — its PR, checks, and unresolved review findings — grounded in live git/GitHub state, not memory. Use when the user says runes, asks to resume, "where were we", "what's the status", or wants to pick a branch back up without re-explaining it.

Have the AI branch, commit, push, and create or update a pull request in the current repository. Use when the user says ship, asks to prepare a branch or commit, or wants help following the repo branch and PR workflow without pushing to the base branch.

Have the AI choose the next orchraft lifecycle step and hand off to the right skill, without performing git or GitHub mutations. Use when the user says warchief, asks what should happen next, or wants orchraft to route the work.

Have the AI draft an implementation plan grounded in this repo's own conventions, policies, and precedent before code gets written. Use when the user says warplan, asks to plan a feature or fix, wants a design pass before implementation, or asks how to approach a change.

Have the AI surface a short read-only status nudge about open PR work without mutating git or GitHub. Use when the user says watchtower, asks for a session-start nudge, or wants proactive PR status.

Have the AI explain something — a PR/diff, a file or module, an error or log, a repo policy, or a dependency/changelog — grounded in real sources (code, git history, docs, policies) with citations, not a confident guess. Use when the user says yap, asks "explain this", "why does this exist", "what changed here", or wants a walkthrough of an error or unfamiliar code.