accenture/mercury/archive-fact
Claude Code and agent skill by accenture/mercury
- community
Checking that the SKILL.md file is still in the repository.
Safely move a decayed fact or completed thread from continuity.md into the archive — the mechanical half of a memory review, done deterministically. Use during the REVIEW.md ritual (step 4) or whenever a specific fact must be archived, instead of hand-editing continuity.md (which has truncated the file twice). The agent decides what to archive; this performs the move safely.
Install
npx vskill@latest install accenture/mercury/archive-factSKILL.md
⚠️ Tool-managed skill provided by agent-memory (provenance: agent-memory-builtin). Don't edit itin place — fork under a new name, or upstream a fix to the agent-memory project (seeSKILLS.md).
This skill is the safe archive-move. It performs REVIEW.md step 4 — moving a faded/superseded fact's block out of the live layer — continuity.md, or a thread's own memory/open-threads/thread-<id>.md file (v4.39.0; the sweep moves the block to the quarter archive and deletes the file) — as a runnable script, so the move can't be botched. Don't hand-edit `continuity.md` to do this (the read-modify-write open(f,"w").write(open(f).read()+…) truncates the file before the read — it has wiped a version.md stamp and then this repo's archive, 50 facts → 6, once each). This script reads the whole file into memory first and writes once, so truncation is structurally impossible.
What it does NOT do
It does not decide what to archive — that judgment (which facts have faded, contradiction scans, supersession) stays with the agent and the REVIEW.md ritual (never-pick-a-winner). This is the mechanical executor only: you name the id(s); it moves them. Same split as memory-lint (the script counts; the agent judges).
How to run
From the repo root, use whichever runtime the machine has (output is byte-identical):
Flags:
--id ID— the fact/thread id to archive (repeatable; the move is all-or-nothing — if any id is
missing it refuses and changes nothing).
--reason—faded(default) or e.g."superseded by <new-id>".--quarter YYYY-Qn— targetmemory/archive/<quarter>.md(default: derived from today, UTC).--note TEXT— theINDEX.mdone-liner (single--idonly; otherwise derived from the block).--dry-run— print what would move; change nothing. Preview first when unsure.
It (1) extracts each id's block (the column-0 - bullet down through its <!-- id: … --> footer, whole), (2) appends it to the quarter archive under a dated heading and adds an INDEX.md line (append mode — never truncates), and (3) rewrites continuity.md without those blocks (read-into-variable, single write).
Always run `memory-lint` afterward (it confirms no id ended up in both places and no link dangles). The helper is safe by construction; the lint is the deterministic proof.
Guards (it refuses, exit 1)
- an id with no footer in
continuity.mdormemory/open-threads/(typo / already moved); - an id already present in the archive (no double-archive);
- a move that would leave
continuity.mdempty.
Notes
- No runtime? Don't hand-roll the truncating write. Fall back to the manual safe method
REVIEW.md
documents (append-mode / read-into-variable), then run memory-lint. (Same stance as memory-lint: install Python or Node rather than improvise the dangerous step.)
- Two runtimes at output parity, with mirror tests (
test_archive_fact.py/.mjs) — the deterministic
guarantee shouldn't depend on which runtime the machine happens to have.
- Superseded facts: pass
--reason "superseded by <new-id>"; mark the supersession links in the footers
(DECAY.md §9) before or after — the helper moves the block as-is.
Security
- Security scanPassed, 100 of 100Pattern checks on the skill's files.
- Model reviewNot runA model review is scheduled for the next rescan.
Quality evaluation
- Task evaluationNot runNo evaluation yet. Run one in Skill Studio.
Source
- SKILL.md in the repositoryCheckingagent-skills/archive-fact/SKILL.md
Popularity
- Repository stars
- 91
- Repository forks
- 40
- 7-day trend score
- 23
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/accenture/mercury/archive-fact)