agent-instructions
Use when writing project instructions for a coding agent (CLAUDE.md, AGENTS.md, or equivalent). Covers what belongs in them, what does not, structure, and…
Use when turning a meeting, call, or transcript into notes. Covers extracting decisions and actions rather than transcribing discussion, and producing something that is useful a month later.
$ npx -y skills add nimadorostkar/Claude-Skills-collection --skill meeting-notes --agent claude-codeHow it fires
How this skill gets triggered: by you, by Claude, or both.
/meeting-notesContext preview
The summary Claude sees to decide when to auto-load this skill.
Use when turning a meeting, call, or transcript into notes. Covers extracting decisions and actions rather than transcribing discussion, and producing something that is useful a month later.
name: meeting-notes description: Use when turning a meeting, call, or transcript into notes. Covers extracting decisions and actions rather than transcribing discussion, and producing something that is useful a month later. metadata: category: productivity version: 1.0.0 tags: [meetings, notes, summarization, transcripts, actions]
Turn a meeting into a record of what was decided and what happens next. A transcript is not notes; nobody reads a transcript, and the value of a meeting evaporates within a week without a record of its decisions.
1. **Extract the decisions first** — What was actually decided? Not discussed, not raised, not considered — decided. This is what the meeting produced. 2. **Extract the actions, with an owner and a date** — An action with no owner will not be done. An action with no date will be done eventually, which means never. 3. **Record the open questions** — What was raised and not resolved? These are what the next meeting is for, and they are routinely lost. 4. **Capture the reasoning behind each decision** — Not the whole debate. Why this option, and what was rejected. In six months, the reasoning is the part people need. 5. **Write for someone who was not there** — They have no context. A note that says "agreed to go with Sam's approach" is useless to them and, in two months, to you. 6. **Send it the same day** — Notes sent a week later are archaeology.
**Notes that are useful in six months:**
# Platform sync — 2026-07-14 Present: Maya (chair), Sam, Aisha, Tom. Apologies: Jo. ## Decisions **1. We will use PostgreSQL advisory locks for job deduplication, not Redis.** Rejected Redis because it adds an operational dependency for a problem that occurs once per night, and we have no Redis in production today. Rejected "designate one instance" because it is a single point of failure. Revisit if we ever need job scheduling without a database. **2. The integration test gate ships Monday 21 July, not this Friday.** Aisha raised that three teams still use `deploy --force` in their scripts. Shipping Friday would break their deploys with no notice. One week's notice, with an announcement going out today. **3. We are NOT migrating to a new CI provider this quarter.** Discussed at length. The cost is 3-4 weeks and the benefit is ~6 minutes per pipeline run. Not worth it now; revisit if pipeline time exceeds 20 minutes. ## Actions | Action | Owner | Due | |-------------------------------------------------------|-------|---------| | Announce the deploy gate change in #engineering | Maya | 14 Jul | | Implement advisory-lock deduplication | Sam | 18 Jul | | Audit and remove `deploy --force` from the 3 repos | Aisha | 18 Jul | | Write the ADR for the locking decision | Sam | 21 Jul | ## Open - **Do we need job scheduling in the workers, or only in the API?** Tom raised this; not resolved. Tom to bring a proposal to the next sync. - Jo was absent and owns the CI budget. Decision 3 assumes no budget pressure to move — Maya to confirm with Jo before Friday.
Someone who was not in the room can read this in ninety seconds and know exactly what happened, what they owe, and what is still open.
A curated library of 137 production-grade skills for Claude and other AI coding agents. Every skill follows one structure, speaks with one voice, and earns its place by changing what the agent does.
Repo: nimadorostkar/Claude-Skills-collection
Use when writing project instructions for a coding agent (CLAUDE.md, AGENTS.md, or equivalent). Covers what belongs in them, what does not, structure, and…
Use when an agent needs state that survives a session or a context compaction. Covers what to persist, file-based memory, structuring notes for retrieval, and…
Use when automating agent behavior with lifecycle hooks. Covers hook events, deterministic enforcement of rules the model should not be trusted to remember,…
Use when packaging skills, commands, hooks, and MCP servers into a distributable plugin. Covers manifest structure, bundling, versioning, testing, and…
Use when writing a new skill for an AI agent. Covers scoping, description writing for reliable triggering, progressive disclosure, and the difference between a…
Use when reviewing or improving an existing agent skill. Covers triggering accuracy, content quality, redundancy with the base model, and measuring whether the…