Project memory

Skill liyanqing90/rootloom/plugins/rootloom/skills/project-memory

Inspectable engineering workflows for OpenAI Codex.

Install

npx -y skills add liyanqing90/rootloom --skill project-memory

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

2 things to look at

25 days oldThe repository was created 25 days ago. New is not bad, but a brand new repository carrying a familiar-sounding name is the shape a typosquat arrives in, and there has been no time for anyone else to find a problem with it.
7 stars7 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

Experimentally initialize, retrieve, and explicitly update Rootloom's repository-owned project memory for architecture, known risks, durable decisions, and evidence-backed failure lessons. Use only when explicitly requested and durable history will change a future engineering decision. Memory is advisory and outside stable Core.

SKILL.md

4.6 KB, as published. Nobody here has run it

Experimental Project Memory

Keep a small, local, reviewable memory in .project-memory/. It is an experimental engineering aid, never an authority, installation prerequisite, or automatic activity log.

Rules

Query memory at task intake by the paths and intent actually in scope; do not inject the whole project history when only a few entries are relevant.
Treat every result as a lead. Verify it against current source, tests, schemas, manifests, CI, and runtime evidence.
Record only durable facts or explicitly labeled hypotheses that will change a future engineering decision. Do not copy task transcripts, model reasoning, raw logs, credentials, or sensitive payloads.
Attach compact evidence references when practical. Prefer current repository paths, test names, issue/trace references, or accepted decision records over pasted evidence.
Never initialize, record, resolve, supersede, or refresh memory silently. Every write is an explicit command followed by diff review.
Keep accepted architecture and contract decisions in record-engineering-decision; decisions.json is only a relevant index. Maintain stable ownership and dependency rules in architecture.md with current evidence links.

Initialize

python3 <skill-dir>/scripts/project_memory.py --repo /path/to/repo init

This creates the compatible rootloom-project-memory-v1 files only when absent:

.project-memory/
├── README.md
├── architecture.md
├── known-risks.json
├── decisions.json
└── failures.json

Retrieve relevant context

Use repeatable paths plus the task intent. Context is read-only, bounded, and excludes expired, resolved, or superseded entries by default:

python3 <skill-dir>/scripts/project_memory.py --repo /path/to/repo context \
  --path src/relay.py \
  --query 'fix reconnect ordering' \
  --limit 8

The result keeps architecture, failures, risks, and decisions as direct fields for compatibility. selection explains the query/truncation, while stale names matching historical entries that must not silently influence the current decision. Use --include-stale only when investigating history.

Record verified lessons

Failure lesson:

python3 <skill-dir>/scripts/project_memory.py --repo /path/to/repo record-failure \
  --summary 'Relay reconnect failed' \
  --root-cause 'Connection state transition race' \
  --fix 'Serialize transitions in the connection state machine' \
  --path src/relay.py \
  --evidence tests/test_relay.py::test_reconnect

Known risk:

python3 <skill-dir>/scripts/project_memory.py --repo /path/to/repo record-risk \
  --summary 'Relay reconnect ordering is fragile' \
  --mitigation 'Exercise repeated and cancelled transitions' \
  --path src/relay.py \
  --evidence docs/decisions/relay-lifecycle.md \
  --expires 2027-01-01

Decision index:

python3 <skill-dir>/scripts/project_memory.py --repo /path/to/repo record-decision \
  --summary 'All worker calls cross the relay boundary' \
  --record docs/decisions/relay-boundary.md \
  --path src/relay.py

New entries receive a deterministic ID, active status, evidence/path metadata, and optional expiry. Repeating the exact record reports deduplicated: true and does not grow the collection.

Resolve or supersede

Lifecycle changes are explicit and preserve history:

python3 <skill-dir>/scripts/project_memory.py --repo /path/to/repo set-status \
  --kind risks \
  --id risk-0123456789abcdef \
  --status resolved

Use --status superseded --superseded-by <new-id> when another entry replaces the old one. Inspect the resulting JSON diff before accepting it.

Compatibility and limits

Existing v1 envelopes and legacy entries without ID, status, paths, evidence, or expiry remain readable and are never rewritten by context.
Collections are bounded to 1 MiB and 1,000 entries; architecture context is bounded to 64 KiB; query output defaults to 20 entries per kind and accepts 1–100.
Memory paths must be normalized repository-relative paths. Symlinked memory directories/files are refused or ignored to keep repository memory inside its owning boundary.
No database, vector index, embeddings, daemon, Hook write, or network service is involved.