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 agentmods add commands/modeled-information-format/mnemonic/custodiangit clone --depth 1 https://github.com/modeled-information-format/mnemonicWhat 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 | $0.00014 | $0.01087 |
| Opus 5 | $0.00007 | $0.00544 |
| Sonnet 5 | $0.00003 | $0.00217 |
| Haiku 4.5 | $0.00001 | $0.00109 |
Grade A, and why
custodian 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 yesterday.
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 — 168 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Memory
Search first: /mnemonic:search {relevant_keywords}
Capture after: /mnemonic:capture {namespace} "{title}"
Run /mnemonic:list --namespaces to see available namespaces from loaded ontologies.
/mnemonic:custodian
Comprehensive memory system custodian for health checks, validation, and maintenance.
Usage
/mnemonic:custodian [operation] [options]
Operations
audit (default)
Full health check: links, decay, relationships, orphans, frontmatter completeness.
/mnemonic:custodian audit [--fix] [--dry-run] [--json]
Runs all checks in a single pass:
- Validate memory frontmatter (required fields, types, formats)
- Validate wiki-links and frontmatter relationship targets
- Validate ontological relationship types
- Recalculate decay strength values
- Detect orphaned memories (no incoming references)
relocate
Move memories when project or org names change. Auto-updates all cross-references.
/mnemonic:custodian relocate <old-path> <new-path> [--dry-run] [--commit]
Example:
# Project renamed from old-api to new-api
/mnemonic:custodian relocate ${MNEMONIC_ROOT}/zircote/old-api ${MNEMONIC_ROOT}/zircote/new-api
Steps:
- Build link index across all memory roots
- Update path references in all memories
- Move files (using
git mvwhen available) - Clean up empty directories
validate-links
Check and optionally repair wiki-links, UUID references, and frontmatter relationships.
/mnemonic:custodian validate-links [--fix] [--json]
With --fix: removes broken [[wiki-links]] from body text (replaces with plain text).
decay
Recalculate exponential decay strength values and update frontmatter in-place.
/mnemonic:custodian decay [--dry-run] [--json]
Formula: strength = current_strength * 0.5^(days_since_access / half_life_days)
Supports models: exponential, linear, step, none.
summarize
Find memories eligible for compression. Reports candidates for the compression-worker agent.
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.
- yesterday First seen · 168 lines · 14 tokens per session scan A 258b3ce3ad33
custodian is a command published in the GitHub repository modeled-information-format/mnemonic (22 stars, last pushed 1mo ago), licensed MIT. It adds 14 tokens to every session and 1,087 once invoked, about $0.0001 per session on Opus 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 2026-08-30.
Other commands, from other repositories
issue-review
Run Codex native + adversarial review against the active issue, scoped to allowedfiles, capped per kind.
wiring-check
End-of-task wiring gate — verify every change is connected end-to-end across kipi plugins, hooks, MCP tools, agents, bus files, canonical, and rules. Nothing dangling.
issue-closeout
Triage Codex findings via per-finding dispositions, record findingstriaged, close the active issue.
prd-personas
Run the Skeptic persona session against the active draft PRD.
linear-drain
Create the queued Linear projects and issues that shell scripts captured offline.
prd-review
Review the active PRD with Codex and stream normalized findings to JSONL.