agentsclimarketplace

Reviewing memory

Skill davidlee/doctrine/plugins/doctrine/skills/reviewing-memory

bathe your agents in engineering rigour and flames

Install
npx -y skills add davidlee/doctrine --skill reviewing-memory

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

  • 1 stars1 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its author says it does

Copied from the file, not written here

Structured audit for stability gates — before releases, migrations, large refactors, or when agent confusion is detected. Use to prune, correct, and re-anchor the memory corpus before it drifts past useful.

SKILL.md

2.8 KB, 618 tokens by cl100k_base, as published. Nobody here has run it

Reviewing Memory

MCP shortcut: If the doctrine MCP server is connected, use memory_show via MCP tool instead of doctrine memory show for machine-parseable JSON results with backlinks.

A stability-gate audit: systematically pull the highest-impact memories, inspect their quality, and produce an auditable outcome for each. Run before a release, a schema migration, an ADR that shifts subsystem boundaries, or when retrieve returns contradictory or low-confidence results.

Procedure

  1. Pull highest-impact memories. Run doctrine memory validate corpus-wide (no REF) to surface dangling relations, stale verifications, and expiring drafts across the entire corpus. The output is your audit queue.

  2. Prioritize. Order findings by impact:

    • Memories with scoped paths that have seen high commit churn since attestation (drift risk)
    • Attested (verified) memories — stale attestation on a moving target is worse than unattested
    • High severity × high weight — wrong here hurts

  3. Run the checklist on each. For every memory in the audit queue, check:

    • Provenance. Is the source still traceable? Has the cited code/doc moved?
    • Freshness. Is the attestation recent? Has the scoped path churned past usefulness?
    • Metadata efficiency. Is the scope still accurate? Are tags stale? Is the lifespan appropriate?
    • Scope accuracy. Do the path/glob/command scopes still cover the right surface?
    • Actionability. Can a future agent act on this without further research?
    • Duplication. Does this memory overlap with another? Merge or supersede.

  4. Thread hygiene. Scan for lingering unverified threads (memory list --type thread). For each:

    • If still relevant, verify it (doctrine memory verify <REF>)
    • If stale, archive it (doctrine memory status <REF> archived)
    • If superseded by newer knowledge, supersede it (doctrine memory status <REF> superseded --by <SUCCESSOR>)

  5. Produce outcomes. Every reviewed item must land in exactly one terminal state:

    • verified — re-attested on current tree, metadata updated
    • corrected — edited (doctrine memory edit) with updated scope, trust, or content
    • superseded — pointed at its successor via memory status superseded --by
    • archived — no longer actionable but kept for history
    • promoted — converted to a durable artifact (ADR, spec, or doc) and archived from memory

    No memory leaves review without one of these outcomes recorded.

Keep looking

Skills are one crate of 328,083. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.