Grounded Vault
Unverified●31/40Claude Code◐PartialHas SKILL.md but declares no allowed-tools — Claude Code will ask for permission each time
Cursor·UnknownWe have not crawled the repo tree, so we will not guess
Codex·UnknownWe have not crawled the repo tree, so we will not guess
Gemini CLI·UnknownThe spec defines no detection rule for Gemini
Copilot·UnknownWe have not crawled the repo tree, so we will not guess
npx agentalley add grounded-vaultWho is stuck, and on what
Use when maintaining a durable Markdown knowledge store that agents compile from sources, when every number or quote in a wiki page must trace back to an immutable source, or when compiled pages need cheap drift detection against the code they describe. Teaches the raw/wiki/archive layout, per-claim provenance links, and git fingerprints for zero-token staleness checks.
The whole source
Frontmatter — 2 properties
| name | grounded-vault |
|---|---|
| description | Use when maintaining a durable Markdown knowledge store that agents compile from sources, when every number or quote in a wiki page must trace back to an immutable source, or when compiled pages need cheap drift detection against the code they describe. Teaches the raw/wiki/archive layout, per-claim provenance links, and git fingerprints for zero-token staleness checks. |
| 1 | --- |
| 2 | name: grounded-vault |
| 3 | description: Use when maintaining a durable Markdown knowledge store that agents compile from sources, when every number or quote in a wiki page must trace back to an immutable source, or when compiled pages need cheap drift detection against the code they describe. Teaches the raw/wiki/archive layout, per-claim provenance links, and git fingerprints for zero-token staleness checks. |
| 4 | ---A5 — No allowed-tools declared — no way to tell what this skill may touch |
| 5 | |
| 6 | # Grounded Vault |
| 7 | |
| 8 | A grounded vault is a three-layer Markdown store in which every compiled claim can be traced |
| 9 | back to an immutable source and every page can be checked for staleness with one `git diff`. |
| 10 | It needs a git repository and nothing else. The convention comes from the `llm-wiki-loop` |
| 11 | project, which is a reference implementation rather than a dependency; this skill teaches the |
| 12 | pattern so it works with plain files and whatever agent is in the session. |
| 13 | |
| 14 | ## When to Use |
| 15 | |
| 16 | - An agent compiles notes, papers, transcripts, logs, or code into wiki pages that later sessions rely on. |
| 17 | - A page cites numbers, dates, or quotes, and a reader must be able to verify each one against its source. |
| 18 | - Pages describe code, and rereading the codebase every session to check whether they still hold is too expensive. |
| 19 | - Knowledge must be corrected without losing history: superseded pages are archived, never deleted. |
| 20 | |
| 21 | Session state, task queues, and conversation continuity are a different problem; use the |
| 22 | context-management or conductor plugins for those. This skill is about provenance and drift on a |
| 23 | durable knowledge store. |
| 24 | |
| 25 | ## The three layers |
| 26 | |
| 27 | | Layer | Contents | Who writes it | Rule | |
| 28 | |---|---|---|---| |
| 29 | | `raw/` | source material: notes, papers, transcripts, logs, exported data | people and ingestion only | immutable once added; agents never edit a raw file | |
| 30 | | `wiki/` | compiled pages built from `raw/` and from code | agents and people | every number, date, and quote links to its source | |
| 31 | | `archive/` | pages that drifted or were superseded | agents, during garbage collection | moved, never deleted; the header says why | |
| 32 | |
| 33 | Two files sit at the vault root. `index.md` is the map of every current page. `log.md` is an |
| 34 | append-only record of what changed and why. Both change in the same commit as the page they |
| 35 | describe. |
| 36 | |
| 37 | ## Page header contract |
| 38 | |
| 39 | Every `wiki/` page opens with a header block: |
| 40 | |
| 41 | ```markdown |
| 42 | # Authentication architecture |
| 43 | |
| 44 | > Raw: [raw/notes/auth-v1.md](../raw/notes/auth-v1.md), [raw/adr/0007-jwt.md](../raw/adr/0007-jwt.md) |
| 45 | > Fingerprint: git:5b237fa |
| 46 | > Monitored: src/auth/jwt.ts, src/auth/session.ts, package.json |
| 47 | > Status: Current |
| 48 | ``` |
| 49 | |
| 50 | - `Raw:` lists every source the page was compiled from. Inline claims link to their specific source as well: `Tokens expire after 15 minutes ([raw/adr/0007-jwt.md](../raw/adr/0007-jwt.md)).` |
| 51 | - `Fingerprint:` is the short commit hash the page was compiled against. |
| 52 | - `Monitored:` lists the code paths the page describes. A change to any of them after the fingerprint means the page may be stale. |
| 53 | - `Status:` is `Current`, `Outdated` (monitored code moved on), or `Disputed` (a newer source contradicts the page). |
| 54 | |
| 55 | ## Grounding rule |
| 56 | |
| 57 | A compiled page states only what a source supports, and every number, date, or quotation |
| 58 | appears verbatim in the linked source. A synthesis says it is one and links its inputs. A gap |
| 59 | in the sources is written into the page as a gap rather than filled by guessing. |
| 60 | |
| 61 | Check it mechanically: for each linked claim, search the linked raw file for the exact figure |
| 62 | or quoted phrase. A miss is a grounding error and blocks the commit. The script in |
| 63 | `references/details.md` does this for a whole vault. |
| 64 | |
| 65 | ## Drift detection |
| 66 | |
| 67 | Compare the fingerprint with the current tree instead of rereading monitored code: |
| 68 | |
| 69 | ```bash |
| 70 | git diff --stat 5b237fa..HEAD -- src/auth/jwt.ts src/auth/session.ts package.json |
| 71 | ``` |
| 72 | |
| 73 | Empty output means the page still describes the code it was compiled against. Any output means |
| 74 | recompile: reread only the changed files, update the page, and stamp the new fingerprint. The |
| 75 | check runs in milliseconds and spends no model tokens. |
| 76 | |
| 77 | ## Workflow |
| 78 | |
| 79 | 1. **Ingest.** Put new material in `raw/` under a dated or sourced filename. Never rewrite an existing raw file; add a new one beside it. |
| 80 | 2. **Compile.** Write or update the `wiki/` page with the header block, a source link on every claim, and the fingerprint of the commit the code was read at. |
| 81 | 3. **Check.** Run the grounding check and the drift check before committing. Fix misses at the source; do not weaken a claim to make the check pass. |
| 82 | 4. **Garbage collect.** When drift or a contradicting source appears and the page is not recompiled now, change its status, move it to `archive/`, and record the reason: |
| 83 | |
| 84 | ```markdown |
| 85 | > Status: Outdated |
| 86 | > Reason: src/auth/session.ts changed after git:5b237fa; see log.md 2026-09-01 |
| 87 | ``` |
| 88 | |
| 89 | 5. **Update the map.** Every add, move, or archive updates `index.md` and appends one line to `log.md` in the same commit. |
| 90 | |
| 91 | ## Commit gate |
| 92 | |
| 93 | Run both checks from a pre-commit hook or a CI step so a page cannot land with an unverifiable |
| 94 | number or a stale fingerprint: |
| 95 | |
| 96 | ```bash |
| 97 | python3 scripts/check_vault.py --strict # exits 1 on any grounding miss or drifted page |
| 98 | ``` |
| 99 | |
| 100 | ## Going deeper |
| 101 | |
| 102 | `references/details.md` covers: the vault check script, batching the drift check across pages, |
| 103 | templates for `index.md` and `log.md`, renamed or deleted monitored files, sources that are |
| 104 | binary or live at external URLs, and the reference implementation. |
| 105 |
Reviews
Installed this one?Write the first review and take the Trailblazer badge.
Alternatives
Research Review via External Reviewer Backend (ultra reasoning)Get a deep critical review of research from an external reviewer backend (Codex or manual). Use when user says "review my research", "help me review", "get external review", or wants critical feedback on research ideas, papers, or experimental results.●····●38/40Training CheckPeriodically check WandB metrics during training to catch problems early (NaN, loss divergence, idle GPUs). Avoids wasting GPU hours on broken runs. Use when training is running and you want automated health checks.●····●37/40Changelog AutomationAutomate changelog generation from commits, PRs, and releases following Keep a Changelog format. Use when setting up release workflows, generating release notes, or standardizing commit conventions.◐····●36/40Hermes Tweet> Install and operate Hermes Tweet, a Hermes Agent plugin for X/Twitter research, timeline reading, tweet analysis, and approval-gated private or state-changing operations. Use this skill when installing Hermes Tweet, researching X/Twitter accounts, monitoring launch signals, investigating mentions, auditing giveaways, or preparing gated X operations. Use proactively when a Hermes Agent workflow needs current X/Twitter context. Requires XQUIK_API_KEY for read and action tools.◐····●36/40