Skip to content

snapsynapse/skill-provenance / skills/open

Verify a skill bundle's integrity when opening a project. Use when loading a skill project directory and you need to check file completeness, verify SHA-256 hashes against MANIFEST.yaml, detect stale files, review recent changes, or understand what state the bundle is in before starting work. Triggers on: "open this skill", "verify the bundle", "check what changed", "is everything up to date", "load this project".

Skill Provenance: Open

Verify a skill bundle when opening a session. This ensures you start from a known-good state before making any changes.

When to use

Run /skill-provenance:open at the start of any session where you'll work on a skill bundle — a directory containing SKILL.md and optionally MANIFEST.yaml, CHANGELOG.md, evals, scripts, and other tracked files.

Protocol

  1. Read MANIFEST.yaml first. It is the single source of truth for what the bundle contains. If no manifest exists, tell the user and suggest running /skill-provenance:bootstrap instead.

  2. Verify the inventory is complete. Report any files listed in the manifest but missing from disk, and any files on disk the manifest does not list. Unlisted files fail verification.

  3. Verify structure and hashes. Prefer the bundle's validate.sh when available so constrained path grammar, duplicate inventory entries, symlink components, missing or malformed hashes, and explicit hash: null opt-outs are handled consistently. Otherwise compute SHA-256 manually for every manifest-listed pinned file and report reduced assurance.

  4. Read CHANGELOG.md. Summarize the most recent 2-3 entries so the user knows what changed recently.

  5. Check for staleness. Use hash drift, recent changelog dependency notes, internal-header conflicts, validated_against, and deployment records. Never compare per-file revision integers numerically with bundle semver; they are separate version domains.

  6. Check deployment drift. If the manifest has a deployments section, note any deployed copies whose version appears behind the current bundle version. Don't alarm — just surface the information.

  7. Resolve conflicts if found. If a file's internal version header disagrees with the manifest's version for that file, or two files claim different bundle versions:

    • Present the specific discrepancy to the user.
    • Show what each version claims via its change_summary.
    • Default recommendation: trust the most recent version_date.
    • Always ask the user for explicit confirmation. Never silently resolve a version conflict.
  8. Report. Give the user a concise summary:

    • Bundle name and version
    • File count: present / expected
    • Hash check: pass / N mismatches
    • Structural errors and explicit hash opt-outs
    • Matching, stale, or malformed attestation records when present
    • Stale files (if any)
    • Deployment drift (if any)
    • Recommended first action for this session

Conventions referenced

This skill follows the Skill Provenance open standard. Key conventions:

  • Version identity lives inside files (via YAML frontmatter metadata block) and in MANIFEST.yaml, never in filenames.
  • MANIFEST.yaml is not self-listed. It tracks other files but does not contain its own hash.
  • Per-file versions are integers counting revisions. Bundle version (bundle_version) is semver.
  • Hashes are SHA-256 of file contents, computed on save, verified on load.

For the full spec: https://skillprovenance.dev

reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis.Stars: 467kUpdated Claude CodeCodexVerifiedInstall
diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness.Stars: 319kUpdated Claude CodeCodexVerifiedInstall
dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependenciesStars: 319kUpdated Claude CodeCodexVerifiedInstall
executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is availableStars: 319kUpdated Claude CodeCodexVerifiedInstall