context

Maintain durable project context in `tasks/context.md` (state, decisions, milestones, gotchas, optional context links), inline during other workflows or standalone for cleanup/backfill. Triggers: update context.md, decision log, record project context, capture high-value reference links that improve context handoff.

Safety Notice

This listing is imported from skills.sh public index metadata. Review upstream SKILL.md and repository scripts before running.

Copy this and send it to your AI assistant to learn

Install skill "context" with this command: npx skills add kelvinz/cobb/kelvinz-cobb-context

context

Maintain tasks/context.md as a living, durable "what matters" record for the project. Write it for someone taking over the project (or a future session resuming later). Explain where things are, what was decided, and what to watch before proceeding. Treat this as shared behaviour embedded in other skills, not a mandatory standalone step in normal workflows. When invoked inline (from another skill), apply this workflow as part of that skill's execution. When invoked directly, use for explicit backfill/repair/cleanup requests.


Guardrails

  • Do not store secrets (API keys, tokens, credentials).
  • Keep entries short and durable; avoid duplicating long PRDs or code.
  • Update tasks/context.md in place; do not rewrite the whole file.
  • Write in plain language so a junior dev (or another AI) can take over without extra context.
  • Prefer inline updates during active skills (prd, design, implement, review, commit).
  • Update when worthwhile information emerges.
  • Links are optional and high-signal only; see Links Format below.
  • When asking users for decisions/clarifications, provide numbered short-reply options (e.g. 1, 2, 3).

What goes where

  • PRD files (tasks/f-##-*.md) = what we intend to do (the spec + progress checklist for one feature)
  • tasks/context.md = project definition, durable decisions, milestones, gotchas, and "where we are now"

What to Record

  • Project gist (what it is, who it's for, success measures) → ## Project
  • Current state (what's done, what's next, what's blocked) → ## Current state
  • Key decisions and durable design decisions (what we chose + why + tradeoffs) → ## Key decisions
  • Completed work (feature IDs and notable outcomes) → ## Completed
  • Notes/gotchas (constraints, pitfalls, conventions, sharp edges) → ## Notes / gotchas
  • Optional context links (canonical docs, API specs, repos) → ## Links (optional)

Avoid:

  • raw meeting transcripts
  • verbose day-by-day logs
  • duplicate content that already lives in PRDs or code

Links Format

  • Only add ## Links (optional) when links materially improve handoff context.
  • Format: [Label](path-or-url) — why this helps future context
  • Prefer durable references (design docs, API specs, runbooks, repos); avoid ephemeral links (temp chats, preview URLs).
  • De-duplicate by updating existing entries instead of adding near-identical links.

Workflow

  1. Ensure tasks/ exists; create tasks/context.md if missing.
    • Use references/context-template.md as the default file template when creating it.
  2. Add short entries (1–3 lines) in the most relevant section.
  3. Prepend new entries at the top of each section (newest-first order).
  4. Prefer referencing stable feature IDs (f-##) and avoid file paths (paths can change after archiving/compaction).
  5. If a link materially improves context, add or update entries in ## Links (optional) using the links format.
  6. If it's unclear what to record, ask one clarifying question with numbered short-reply options (e.g. 1, 2, 3).

References

  • references/context-template.md: default structure for creating tasks/context.md.

Output

  • Create or update tasks/context.md.
  • Inline (from another skill): reply with a short summary of what was added/updated. Let the calling skill's output format take precedence.
  • Standalone (direct invocation): reply with the file path, summary, and a status block:
    • Files changed: list of created/updated files
    • Key decisions: any assumptions or choices made (if any)
    • Next step: recommended next skill or action

Source Transparency

This detail page is rendered from real SKILL.md content. Trust labels are metadata-based hints, not a safety guarantee.

Related Skills

Related by shared tags or category signals.

General

compact

No summary provided by upstream source.

Repository SourceNeeds Review
General

commit

No summary provided by upstream source.

Repository SourceNeeds Review
General

implement

No summary provided by upstream source.

Repository SourceNeeds Review
General

review

No summary provided by upstream source.

Repository SourceNeeds Review