aidlc-io/aidlc/assets
Claude Code and agent skill by aidlc-io/aidlc
- community
Checking that the SKILL.md file is still in the repository.
Load and maintain an epic's compact memory so continuing it (with any agent) is cheap on tokens. Reads docs/epics/<epic>/epic-memory.json FIRST as primary context, appends durable decisions/constraints as you work, and records a reflection on how to prompt/work better next time. Invoke /epic-context <epic> when starting or continuing an epic.
Install
npx vskill@latest install aidlc-io/aidlc/assetsSKILL.md
/epic-context — cheap, portable epic memory
Every epic can carry a compact memory at docs/epics/<epic>/epic-memory.json:
summary— one paragraph: what the epic is + where it standsentries[]— distilled decisions / constraints / context / notesreflections[]— lessons on how to prompt/work more effectively next time
The point: a future session (any agent) reads this digest FIRST instead of re-reading every artifact + git history — far fewer tokens. So keep it compact and high-signal; it is a digest, not a transcript.
The tool ships with the AIDLC extension at ~/.claude/tools/epic-memory.mjs and runs with just node. Let M = node "$HOME/.claude/tools/epic-memory.mjs". Each write is auto-attributed (git identity, hostname fallback) + timestamped.
Arguments
/epic-context <epic> — e.g. /epic-context EPIC-001. Paths are relative to the open project: EPIC_DIR = docs/epics/<epic>.
On start (continuing an epic)
- Load memory first:
$M show "$EPIC_DIR".
- If it prints a digest, treat it as your primary context. Only open the full
artifacts ($EPIC_DIR/artifacts/*.md) for the specific parts the memory doesn't already cover — don't re-read everything.
- If it says "no epic memory yet", fall back to reading the artifacts, and
seed a summary once you understand the epic: $M summary "$EPIC_DIR" --text "…".
While working
Append durable, reusable facts as you make them (not chit-chat):
$M add "$EPIC_DIR" --kind decision --text "Chose X over Y because …"$M add "$EPIC_DIR" --kind constraint --text "Must not touch Z / must keep …"$M add "$EPIC_DIR" --kind context --text "Key fact a newcomer would need"
Keep the summary current when the epic's state changes materially: $M summary "$EPIC_DIR" --text "…".
At the end (or when the user gives notable feedback)
Record a reflection so the next session prompts/works better: $M reflect "$EPIC_DIR" --text "Next time: state target files + expected result up front to avoid back-and-forth".
Guardrails
- Compact and high-signal — this file exists to save tokens, so don't dump raw
prompts or long transcripts into it.
- It complements, never replaces, the
.mdartifacts (still canonical). - Safe to commit: it travels with the epic so any teammate/agent benefits.
Security
- Security scanPassed, 95 of 100Pattern checks on the skill's files.
- Model reviewPassed
Quality evaluation
- Task evaluationNot runNo evaluation yet. Run one in Skill Studio.
Source
- SKILL.md in the repositoryCheckingpackages/extension/assets/epic-context.skill.md
Popularity
- Repository stars
- 63
- Repository forks
- 30
- 7-day trend score
- 21
Stars and forks are for the whole GitHub repository, not this skill alone. Not refreshed yet.
Works with
Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.
- Claude Code
- Cursor
- GitHub Copilot
- Windsurf
- Codex
- Gemini CLI
- Cline
- Amp
- Roo Code
- Goose
- Continue
- Kilo Code
- Junie
- Trae
- OpenHands
- Qwen Code
- AI Agent
- Tabnine Chat
- Sourcegraph Cody
- Amazon Q
- Firebase Genkit
- Sweep
- Devin
- Phind
- Mentat
- Claude Desktop
- Aider
- GPT Pilot
- Plandex
- bolt.new
- v0
- Replit Agent
- Anse
- TypingMind
- BetterChatGPT
- LibreChat
- Warp AI
- Pieces
- Fig AI
Embed badge
[](https://verified-skill.com/skills/aidlc-io/aidlc/assets)