Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add cimo-labs/kvault --skill kvault-maintenancegit clone --depth 1 https://github.com/cimo-labs/kvaultWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/cimo-labs/kvault/kvault-maintenance)<a href="https://agentmods.dev/skills/cimo-labs/kvault/kvault-maintenance"><img src="https://agentmods.dev/badge/skills/cimo-labs/kvault/kvault-maintenance/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/cimo-labs/kvault/kvault-maintenance"><img src="https://agentmods.dev/badge/skills/cimo-labs/kvault/kvault-maintenance.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00059 | $0.03349 |
| Opus 5.5 | $0.00024 | $0.01340 |
| Sonnet 5.5 | $0.00012 | $0.00670 |
| Haiku 4.5 | $0.00006 | $0.00335 |
Grade A, and why
kvault-maintenance scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 26d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 209 lines — stays where its author put it; the contents beside it link to each section on GitHub.
kvault maintenance — cadence and procedure
A KB rots in a specific way under multi-agent writes: flat parents with
dozens of children, near-duplicate siblings (infra/ beside
infrastructure/), root categories minted one per sub-problem, directories
with no summary that no surface can see, and two journal layouts. kvault check names each of these; kvault plan orders the fixes and emits the
commands. This skill is the procedure around those two commands. The plan
comes from kvault, not from you, so every runtime makes the same moves.
Requires knowledgevault 0.17.0 or later. The first line of every unattended
job is kvault doctor, so a runtime older than the skill text is visible in
the job log instead of silently missing the signals below.
Nothing here waits on a person. Agents decide from the summaries and act;
when unsure they take the reversible action (nest, do not merge; move, do
not delete). The human's part is correcting the agent later, in use, and a
correction is recorded in the tree with kvault mark so it sticks.
Vocabulary
kvault check prints one bounded group per prefix. Hard findings (exit 1)
are on the [KB] line; everything else is warn-only maintenance work.
| Prefix | Meaning | What to do |
|---|---|---|
[KB] PROPAGATE / LOG / WRITE / BRANCH |
Stale parent, missing journal entry, missing frontmatter, parent over the child ceiling | Fix before other work; BRANCH → kvault plan <path> |
SUMMARY: |
A parent rollup is too short, misses children, has placeholder text, is too long, or accretes dated sections | too_short/missing_child_coverage/placeholder_language: rewrite as a rollup. too_long/stale_history: fold, never split — chronology goes to journal/, detail to <node>/deep_context/ |
GHOST: |
A directory with no _summary.md: invisible to tree, search, and check |
Write a summary (kvault write <path> --create) or list it in .kvaultignore if it is tooling |
SERIES: |
A parent whose children differ only by date/time words — a chronology written as nodes (daily cards) | Run the series item from kvault plan: its batch moves the dated nodes under one hub's deep_context/; then write the hub as the current state (update it if new_parent_exists). New timeline entries go to journal/. Never add another dated node to a series |
SIBLINGS: |
Two names under one parent share their words | Same thing → merge and delete one. Subtopic → kvault move. Undecidable → leave it for the monthly review |
DUPLICATE: |
The same thing in two places anywhere in the KB: same name at two depths, same title, shared aliases, or near-identical text | Read both. Same thing → fold the unique facts into one node, then park the other under its deep_context/ (the duplicate item's move; one move undoes it). Different things → kvault mark <a> --distinct-from <b> |
DANGLING: |
A summary links to, or lists as a child, a path with nothing there | Point it at the node's current path (the finding names same-name nodes elsewhere) or drop it |
LOOSE: |
A file outside the node convention. legacy_node_file = Markdown with frontmatter that search cannot see; supporting_doc; artifact |
Adopt legacy node files as nodes (plan emits kvault write <node> --create < file && git rm file); supporting docs into <node>/deep_context/; artifacts ignored |
JOURNAL: |
Files off the journal/YYYY-MM/log.md layout, or a second history |
Fold into the canonical log with kvault journal, then remove. A second layout the owner keeps on purpose (a weekly journal tree) goes in .kvaultignore instead |
STALE: |
A node's verify_by date passed: it records facts that go stale |
Re-check them and rewrite what changed; still time-sensitive → kvault mark <path> --verify-by +14d; settled → --verify-by none |
PENDING: / RETRACTED: |
Captured events never promoted; nodes citing retracted events | Promote or resolve; rewrite and re-link |
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 26d ago Changed · +4 lines 7aa160be9de4
- 1mo ago Changed · +15 lines 9fcc2daf1759
- 1mo ago First seen · 190 lines · 59 tokens per session scan A 8aad87e495f7
kvault-maintenance is a skill published in the GitHub repository cimo-labs/kvault (5 stars, last pushed 27d ago), licensed MIT. It adds 59 tokens to every session and 3,349 once invoked, about $0.0002 per session on Opus 5.5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-11.
Other skills, from other repositories
knowledge-ingest
Ingest documents, articles, conversations, or supplied text into Knowledge with retained originals, sourced claims, entity updates, timelines, and cross-links; also learn durable lessons from a set of conversations, transcripts, or past sessions (/learn), including the Learn from conversations and Learn from Memory…
forget
Remove a personal fact, preference, relationship detail, topic, or prior event from Muse's active memory and stop existing copies or automations from bringing it back. Use for explicit requests such as 'forget that', 'don't remember this about me', or 'remove that from your memory'. Do not use when 'forget it' merely…
mem0-oss-to-platform
Plan, then execute, a migration of a project from the mem0 open-source / self-hosted SDK (local Memory class) to the mem0 Platform / hosted SDK (MemoryClient). TRIGGER when: the developer wants to move, switch, or migrate mem0 off OSS/self-hosted to the hosted API, e.g. "migrate my mem0 setup to the platform", "switch…
Cortex
LifeOS memory, prior-session recall, and learning capture. USE WHEN capture this learning, document this learning, what did we learn recently, recent learnings, cortex, knowledge, knowledge base, search knowledge, what do we know about, archive, harvest, knowledge status, develop note, add to knowledge, add a people…
memory
Use when the user asks to remember, recall, forget, update, search, or inspect durable OpenSquilla memory, including profile facts in USER.md and long-term notes in MEMORY.md or memory//.md.
ha-data-stores
Map of Hope Agent's local data stores and safe read-only query workflow. Use when the user asks where Hope Agent stores data, wants to inspect sessions/messages/memory/logs/background jobs/knowledge indexes/settings, asks the model to query local app data, or debugging requires checking persisted state. Trigger…