agentleFS
karloowskarloows/orchraft/skills/reckoning/SKILL.md

reckoning

Skill1 starsChanged 51 days agoApache-2.0View on GitHub

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.

1 other project uses a version of this.

Adopted by

Most recently taken up by

  • Other versions1 project uses a version of this.

    Word for word · 0

    None found.

    Edited · 0

    None found.

    Wrote their own version · 1

    1. IamK77/Skill★ 2same name

    How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.

  • In this repository19 other files it gives its agents.

The file

Show its rules and text

We split this skill into the rules it gives an agent

17 rules. About 0 are common, 13 are distinctive, and 4 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈13

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 4

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Use this skill when the user wants to see the repo's standing debt of consciously declined review findings — not the current branch's …

    line 8
  2. Very specific:

    Search the repo's pull requests for review-comment replies carrying the <!-- orchraft:declined --> marker ship posts (see …

    line 28
  3. Distinctive:

    Stop. Never touch a PR/issue, and never resolve or reply to anything — this skill only reads.

    line 34
  4. Distinctive:

    Not a backfill tool. Only replies carrying the marker are found.

    line 44
  5. Distinctive:

    Not permission to resolve, reply to, or otherwise act on anything it finds — reporting a debt is not authorization to settle it.

    line 47
  6. Distinctive:

    Run when the user explicitly asks for a reckoning, what's been declined, a debt list, deferred findings, or wants a pre-release check of …

    line 52
  7. Very specific:

    Do not run automatically; every other status-reporting skill here (runes, watchtower) is invoke-on-purpose, and this one is no different.

    line 55
  8. Distinctive:

    Search for marked decline replies across all PR states.

    line 115
  9. Distinctive:

    Keep it to what was actually found — don't pad an empty result with filler.

    line 128
  10. Distinctive:

    Never touch a PR or issue, resolve a thread, or post a reply as part of this skill — it only reads.

    line 140
  11. Distinctive:

    Never report a resolved thread as standing debt just because it once carried the marker.

    line 142
  12. Distinctive:

    Never invent a decline reason that isn't in the actual reply text.

    line 144
  13. Distinctive:

    Never treat an unmarked reply as a declined finding, even if its prose reads like one — the marker is the only signal this skill trusts.

    line 145
  14. Very specific:

    Stop and say so if neither the GitHub MCP connector nor gh is available — don't report a partial or guessed list as complete.

    line 150
  15. Distinctive:

    Stop and say so if any individual repository search, paginated pull-request request, or thread request fails partway through — a search …

    line 152
  16. Very specific:

    Stop after presenting the summary.

    line 159
  17. Distinctive:

    If blocked, drop the orc voice and state the blocker plainly.

    line 170

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Reckoning Workflow

Use this skill when the user wants to see the repo's standing debt of consciously declined review findings — not the current branch's status, the whole repo's.

Contents

  • At A Glance
  • What This Is Not
  • Trigger Rules
  • Requirements
  • Prerequisites
  • Default Path
  • Summary Format
  • Guardrails
  • Stop Conditions
  • Handoff
  • Response Examples

At A Glance

  1. Search the repo's pull requests for review-comment replies carrying the <!-- orchraft:declined --> marker ship posts (see skills/ship/SKILL.md Default Path step 12).
  2. For each one, report the PR, the file/line, the original finding, and the stated decline reason, and whether that PR is still open or already merged/closed.
  3. Stop. Never touch a PR/issue, and never resolve or reply to anything — this skill only reads.

What This Is Not

  • Not runes: runes reports one branch's live status; reckoning reports standing decline debt across the whole repo's PR history.
  • Not roast: it produces no new findings and posts nothing — it reads findings roast (or a human reviewer) already raised and that were already declined via ship.
  • Not a backfill tool. Only replies carrying the marker are found; a decline reply posted before the marker existed won't show up, and this skill never guesses which unmarked replies might have been declines.
  • Not permission to resolve, reply to, or otherwise act on anything it finds — reporting a debt is not authorization to settle it.

Trigger Rules

  • Run when the user explicitly asks for a reckoning, what's been declined, a debt list, deferred findings, or wants a pre-release check of standing debt.
  • Do not run automatically; every other status-reporting skill here (runes, watchtower) is invoke-on-purpose, and this one is no different.

Requirements

  • A GitHub MCP connector configured for this session with at least read access to pull requests and their review comments, or gh installed and authenticated as a fallback. Without either, stop and say the search can't run rather than reporting an empty or partial list as complete.

Prerequisites

  • Confirm which repo to search: the current one by default, or one the user names.
  • Enumerate every pull request in that repo — list_pull_requests with state: all, paginating through every page rather than stopping at the first (a single page defaults to fewer than all of a repo's PRs). Falling back to gh api --paginate "repos/$owner/$repo/pulls?state=all" --jq '.[] | {number, html_url, state, merged_at}' when MCP is unavailable — GitHub reports both merged and unmerged pull requests as state: "closed", so merged_at is what actually distinguishes "merged" from "closed without merging" in this fallback path. gh api --paginate follows GitHub's own pagination until exhausted, unlike gh pr list, whose --limit is always some fixed number and would under-cover a repo whose PR count happens to exceed it. This must cover open, merged, and closed without merging alike, since debt survives a merge.
  • For each PR, fetch its review threads via pull_request_read method get_review_comments, which returns each thread's is_resolved status and its full comment list (original finding plus any replies, each with author/created_at/path/line) in one call, capped at 100 comments per thread — paginate past the first page if a PR has more threads than one page returns. If a thread's returned comment count looks truncated relative to its total (fewer comments came back than the thread actually holds), don't treat that thread as fully read: fall back to the gh api graphql path below for that thread's own paginated comment loop instead, or report it as incomplete rather than risk missing a marker reply past the 100-comment cap. That fallback is a gh api graphql query for the PR's reviewThreads(first: 100, after: $cursor), requesting pageInfo { hasNextPage endCursor } plus each thread's isResolved and comments(first: 100, after: $commentCursor) { pageInfo { hasNextPage endCursor } nodes { body author { login } createdAt path line originalLine } } — loop both the thread page and each thread's own comment page until hasNextPage is false, when the MCP connector is unavailable. Both GraphQL connections cap at 100 items per page; a PR with more threads, or a thread with more comments, needs the loop to actually run, not just the query written once.
  • Within each thread's comment list, find any reply containing the <!-- orchraft:declined --> marker; its thread's other comments hold the original finding text, author, file/line, and timestamp, and the thread's is_resolved field is the settled/standing check — no separate lookup needed for any of these. A comment's current line can be null (an outdated comment whose line no longer maps to the current diff); use original_line/originalLine instead when that happens, and report line unavailable only if neither coordinate exists.

Default Path

  1. Resolve the target repo (see Prerequisites).
  2. Search for marked decline replies across all PR states.
  3. For each one still on an unresolved thread, record: PR number/URL and state (open, merged, or closed without merging), file and line, the original finding (severity if stated), the decline reason, and who declined it.
  4. Drop any match whose thread was later resolved — the marker means it was declined at the time, not that it must stay open forever; a human resolving it later is a legitimate settling of that debt.
  5. Group by PR, most recently declined first.
  6. Report the list, or that none exist.

Summary Format

Keep it to what was actually found — don't pad an empty result with filler.

  • Repo — name, and how many PRs were searched (open, merged, and closed-without-merging counts, mutually exclusive).
  • Standing debt — one entry per unresolved declined finding: PR number/URL and state, file:line, the original finding, the decline reason, and who declined it.
  • None found — say so plainly if the search turned up nothing, and distinguish that from "couldn't search" (see Requirements).

Guardrails

  • Never touch a PR or issue, resolve a thread, or post a reply as part of this skill — it only reads.
  • Never report a resolved thread as standing debt just because it once carried the marker.
  • Never invent a decline reason that isn't in the actual reply text.
  • Never treat an unmarked reply as a declined finding, even if its prose reads like one — the marker is the only signal this skill trusts.

Stop Conditions

  • Stop and say so if neither the GitHub MCP connector nor gh is available — don't report a partial or guessed list as complete.
  • Stop and say so if any individual repository search, paginated pull-request request, or thread request fails partway through — a search that started with GitHub access available can still fail mid-run (a single page erroring, rate limiting, a transient network fault). Report which part failed and that the result is incomplete; never present a partial list gathered before the failure as the full standing debt.
  • Stop after presenting the summary; do not proceed into ship, roast, or any mutation without the user separately asking for it.

Handoff

  • Unless the user asked for plain output, open with a fresh one-line phrase in the orc voice, speaking as the Reckoner from context/personality.md (the target repo's copy when it exists, otherwise ${CLAUDE_PLUGIN_ROOT}/context/personality.md). If they did, open with the plain summary instead.
  • Present the summary per Summary Format.
  • If blocked, drop the orc voice and state the blocker plainly.

Response Examples

Opening lines are samples of the orc voice; write a fresh one each time.

Standing debt found (illustrative — actual output names real PRs/files):

🗂️ Hrrm. Two debts still stand on the books. ✨

Repo: <owner>/<repo> (N PRs searched: X open, Y merged, Z closed without
merging)
- PR #<n> (<state>): `<path>:<line>` — <one-line finding summary>;
  declined by <decliner> because <the stated reason from the reply itself>.
- PR #<n> (<state>): `<path>:<line>` — <one-line finding summary>;
  declined by <decliner> because <the stated reason from the reply itself>.

Clean books:

🗂️ Hah! The books read clean — nothing declined stands open.

Repo: <owner>/<repo> (N PRs searched: X open, Y merged, Z closed without
merging)

Blocked, no GitHub access:

⚠️ Reckoning is blocked: neither the GitHub MCP connector nor `gh` is
available. No search was performed.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.