harrylettering/codememory/codememory-mark-task

Claude Code and agent skill by harrylettering/codememory

Trust tier 3
  • community
Source check pending

Checking that the SKILL.md file is still in the repository.

Use this skill to durably record the current task or goal so future turns (and future sessions) can recall what we are trying to accomplish. Invoke when the user states a non-trivial objective, when scope is clarified after ambiguity, or when a goal is revised mid-session. Do NOT invoke for trivial follow-ups, micro-edits, or work already captured by an earlier task mark in the same session.

Install

npx vskill@latest install harrylettering/codememory/codememory-mark-task

SKILL.md

CodeMemory — Mark Task

Persists a single durable task / goal into the CodeMemory memory store via the per-session daemon. The result is a Memory Node tagged task + requirement.

When to Use

Invoke this skill when, in the current turn, the user (or you, after clarification) committed to a non-trivial goal that should outlive compaction. Good signals:

  • The user opened the session with a multi-step ask: "implement X across A, B, C".
  • A vague request was sharpened into a concrete objective.
  • The user revised the goal partway through ("actually, we also need to migrate Y").
  • You and the user agreed on acceptance criteria.

Do NOT invoke for:

  • Single-edit asks where the goal is fully captured by the file diff.
  • Conversational pleasantries or scoping questions.
  • Goals already marked earlier in this session — supersede only if the goal materially changed.

How to Invoke

Run the wrapper script with requirement as the endpoint and a JSON payload. Required fields: kind: "task", requirement. Optional: details, acceptance_criteria (array), supersedesNodeId, sourceToolUseId.

bash
CLAUDE_SESSION_ID=${CLAUDE_SESSION_ID} ${CLAUDE_PLUGIN_ROOT}/hooks/scripts/codememory-mark.sh requirement "$(cat <<'JSON'
{
"kind": "task",
"requirement": "Migrate the JSONL watcher to write through the new conversation_store API",
"details": "Old path went through ConversationIngestor; new path exposes idempotent insert keyed by sessionId+sequence.",
"acceptance_criteria": [
"All existing tests pass",
"No direct sqlite writes remain in jsonl-watcher.ts",
"New code path handles JSONL replay idempotently"
]
}
JSON
)"

Run the command exactly as written: Claude Code fills in ${CLAUDE_SESSION_ID} and ${CLAUDE_PLUGIN_ROOT}, and the session id is how the mark reaches this session's daemon and no other. On success it returns JSON like {"ok":true,"conversationId":...,"memoryNodeId":"requirement-task-tool-..."}.

Authoring Guidance

  • `requirement`: one durable sentence describing the goal. Imperative voice.
  • `details`: one short paragraph of supporting context. Skip if the requirement is self-explanatory.
  • `acceptance_criteria`: bullet checks that define "done". Each one short and verifiable.
  • `supersedesNodeId`: set when a previous task node is being replaced — keeps lifecycle clean.
  • `sourceToolUseId`: pass the originating tool_use id when known so retries collapse onto the same node.

After invoking, continue the work — do not summarize the task back at the user.

Security

  • Security scanPassed, 90 of 100Pattern checks on the skill's files.
  • Model reviewPassed

Quality evaluation

  • Task evaluationNot runNo evaluation yet. Run one in Skill Studio.

Source

Popularity

Repository stars
133
Repository forks
11
7-day trend score
24

Stars and forks are for the whole GitHub repository, not this skill alone. Refreshed 18 May 2026, 5 months ago.

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
vskill VERIFIED badge
[![vskill VERIFIED](https://verified-skill.com/api/v1/skills/harrylettering/codememory/codememory-mark-task/badge)](https://verified-skill.com/skills/harrylettering/codememory/codememory-mark-task)

Something wrong with this skill or its page? Tell us.