agentleFS
QuirkybrainQuirkybrain/claude-code-swe-framework/.claude/skills/framework-resume/SKILL.md

framework-resume

Skill1 starsChanged 53 days agoMITView on GitHub

Safely resume framework work by reading the latest Checkpoint, validating repository and Artifact state, detecting drift, revalidating affected work, and recomputing the Minimum Necessary Path. Use after session restart, context loss, pause, or handoff.

The file

We split this skill into the rules it gives an agent

10 rules. About 0 are common, 7 are distinctive, and 3 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈7

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 3

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Run at session recovery, after PAUSED_AT_STAGE, after a handoff/context reset, or whenever repository activity may have occurred since the …

    line 14
  2. Distinctive:

    Do not mutate project work until reconciliation identifies a safe next action.

    line 25
  3. Distinctive:

    Read Project Configuration, latest Checkpoint, linked Artifacts, State, and target.

    line 29
  4. Very specific:

    Run Change Impact for every material drift and mark affected Artifacts STALE/UNVERIFIED and Components UNVERIFIED as appropriate.

    line 34
  5. Distinctive:

    Write a reconciled Checkpoint before continuing substantial work.

    line 39
  6. Distinctive:

    Do not rerun completed work whose evidence remains current and sufficient.

    line 44
  7. Distinctive:

    Do not continue a previously ready node if its dependencies changed.

    line 45
  8. Distinctive:

    No trustworthy Checkpoint exists.

    line 58
  9. Distinctive:

    Run Project Intake when no usable snapshot exists.

    line 65
  10. Very specific:

    Stop when current state is reconciled and a safe next action is identified, when the target remains intentionally paused, or when …

    line 69

≈ “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

Framework Resume

Purpose

Continue from persisted evidence without blindly trusting an old Checkpoint or hidden conversation history.

Trigger

Run at session recovery, after PAUSED_AT_STAGE, after a handoff/context reset, or whenever repository activity may have occurred since the last Checkpoint.

Inputs

  • config/project.md, user request, latest Checkpoint, State, Artifact Registry, and dependency/traceability records.
  • Current repository revision, branch/worktree, working-tree status, files, and available validation evidence.
  • Open clarifications, decisions, assumptions, risks, and prior Minimum Necessary Path.

Prerequisites

  • Locate the latest relevant Checkpoint by ID/time/repository lineage, not filename guess alone.
  • Do not mutate project work until reconciliation identifies a safe next action.

Procedure

  1. Read Project Configuration, latest Checkpoint, linked Artifacts, State, and target.
  2. Inspect current repository revision, branch/worktree, dirty state, and relevant file changes.
  3. Compare repository/Artifact/Component revisions with the Checkpoint.
  4. Detect added, removed, modified, or externally changed assets since the snapshot.
  5. Reopen unresolved Clarifications, assumptions, decisions, risks, failures, and blockers.
  6. Run Change Impact for every material drift and mark affected Artifacts STALE/UNVERIFIED and Components UNVERIFIED as appropriate.
  7. Revalidate affected Artifacts and current Gate evidence; preserve unaffected evidence with rationale.
  8. Reassess Workflow Level if scope/risk changed.
  9. Rerun Gap Analysis and recompute the Minimum Necessary Path from current evidence.
  10. Select the next dependency-ready action or report the exact blocker.
  11. Write a reconciled Checkpoint before continuing substantial work.

Decision Rules

  • Repository evidence and exact revisions outrank stale Checkpoint claims, but conflicts must be investigated rather than silently overwritten.
  • Do not rerun completed work whose evidence remains current and sufficient.
  • Do not continue a previously ready node if its dependencies changed.
  • A prior PASS is invalid when its evaluated revision or basis changed.
  • Reaching a previous Target Stage remains paused unless the user requests a new target.

Outputs

  • Resume/Reconciliation Report with Checkpoint and current repository comparison.
  • Changed-since-checkpoint set and resulting state invalidations.
  • Revalidation results, preserved evidence rationale, updated risks, and reconciled Checkpoint.
  • Recomputed Minimum Necessary Path and next ready action/blocker.

Failure Cases

  • No trustworthy Checkpoint exists.
  • Multiple divergent worktrees/revisions claim continuity.
  • Referenced Artifacts/evidence are missing.
  • Repository contains unattributed or conflicting changes.

Escalation

Run Project Intake when no usable snapshot exists. Use Clarification Gate for ownership/intent conflicts, Failure Diagnosis for unexplained drift, and request explicit merge/rollback authority for divergent work.

Stop Conditions

Stop when current state is reconciled and a safe next action is identified, when the target remains intentionally paused, or when conflict/access/authority blocks recovery. Never continue solely because the old Checkpoint said a node was ready.

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.