Team memory capture
Skill meharajM/context-machine/plugins/team-project-memory/skills/team-memory-capture
Token-efficient shared project memory skills for developers and AI coding agents.
npx -y skills add meharajM/context-machine --skill team-memory-captureAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
One thing to look at
- 0 stars0 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
Capture a short reusable project learning after a non-obvious fix, failed approach, setup requirement, decision, or debugging path is verified and likely to save another agent time or tokens.
The file declares its own license as MIT. That is the author’s claim about this one file, and it is not the same thing as the license GitHub reports for the repository, which is listed with the other numbers below.
SKILL.md
1.7 KB, 343 tokens by cl100k_base, as published. Nobody here has run it
Team Memory Capture
Capture only knowledge likely to prevent meaningful repeated work.
Before Writing
- Search with
$team-memory-searchfor the same signature. - Skip generic knowledge, obvious docs, raw transcripts/logs, temporary state, and duplicates.
- Redact secrets, tokens, personal paths, customer/personal data, private URLs, and exploit details.
- Keep the entry under 250 words; keep
Reuseunder 80.
Status
candidate: plausible but not reproduced.validated: confirmed by test, CI, reproducible command, or user confirmation.stale: no longer preferred.contradicted: false for its stated scope.
Agent observation alone remains candidate.
Format
---
id: <YYYYMMDD-HHMM-short-slug>
status: candidate | validated | stale | contradicted
confidence: low | medium | high
scope: <repo/path/component/tool/runtime>
signature: <normalized searchable signature>
environment: <important versions or none>
last_validated: <YYYY-MM-DD or unknown>
supersedes: []
contradicts: []
---
# <Short title>
## Reuse
<First action, prerequisites, and limits.>
## Resolution
<Smallest useful cause and solution.>
## Verify
`<command/check>`
## Avoid
<Repeatable dead ends or none.>
## Evidence
<Brief test, CI, PR, commit, or confirmation.>
Publish
Default to a draft. Do not commit, push, or open a PR without explicit permission or a repo policy that allows it. Prefer one file per learning.