Skip to content

CruxExperts/LocalSetup / ls/skills/ls-herdr-provenance-drift

ls-herdr-provenance-drift

Verified

Check Herdr source, skill, schema, and release drift; maintain the master provenance manifest and reviewed compatibility without automatically upgrading hosts. Start with ls-herdr-task-routing.

Herdr provenance maintenance

Read the master first if needed: ls-herdr-task-routing

Use localsetup herdr provenance --help. The master owns the single canonical manifest.

Check before updating

Check preserved source hashes offline first. Compare upstream stable release metadata and the manifest's bounded source-watch paths when network evidence is needed. source_watch.commit pins the reviewed watched-source baseline independently of source.commit, which remains the tagged release pin; older manifests without source_watch.commit use the release commit as their legacy baseline. Unchanged version numbers do not prove unchanged skill or API source. The scheduled daily watcher is read-only for repository contents and host state. It compares only the canonical manifest's official release and watched-file evidence. A confirmed newer stable release needs an immutable commit SHA; changed watched bytes are reported independently, even if another path is unknown. The workflow opens or updates a deduplicated provenance-refresh issue handoff for findings and emits a concise run summary. Unknown evidence means unknown freshness, not "current." A handoff gives no authority to upgrade a host. QC Patrol remains a separate workflow.

Keep the reviewed source baseline and tagged release separate from imported skill/schema hashes, installed client version, bundled client schema, and live server version/protocol/endpoint. Never turn a client schema into a claim about every remote server. Compare operations actually used; a newer compatible release need not block them.

Accept one coherent update

  1. Resolve the official release to its full immutable commit and inspect release notes, upstream skill changes, license, and consumed API changes.
  2. Update the preserved upstream evidence and source manifest together. Preserve original notices; LocalSetup's adaptation does not relicense upstream work.
  3. Map changed guidance to the relevant specialist and shared core. Extend the existing owner rather than duplicating commands or transport logic.
  4. Validate affected contracts and realistic tasks, update compatibility evidence and owning generated catalogs, and review the material diff.
  5. Create the normal scoped local checkpoint. Publishing and actual host upgrades remain separate operations with their existing authority.

Do not replace source pins merely because a network check found a newer tag. Do not rewrite historical release evidence or claim an untested version is supported. Apply dependency verification once for a newly selected artifact and reuse applicable evidence; do not turn it into repetitive user prompts.

Live host upgrades route to remote.

Reviewed layout appearance change

The watched-source baseline at 5608de0 changes two layout pane-creation call sites to use the shared appearance helper during a pending host-theme transition. The watched skill, generated API schema, schema definitions and operation definitions are byte-identical to the prior baseline. This bounded review requires no LocalSetup API change. The tagged release, imported evidence and binary pins remain unchanged; source review does not qualify a running client or server.

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: 471kUpdated 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: 322kUpdated Claude CodeCodexVerifiedInstall
dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependenciesStars: 322kUpdated 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: 322kUpdated Claude CodeCodexVerifiedInstall