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 odere-pro/claude-wiki-pages-plugin --skill maintain-contractgit clone --depth 1 https://github.com/odere-pro/claude-wiki-pages-pluginWrote 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/odere-pro/claude-wiki-pages-plugin/maintain-contract)<a href="https://agentmods.dev/skills/odere-pro/claude-wiki-pages-plugin/maintain-contract"><img src="https://agentmods.dev/badge/skills/odere-pro/claude-wiki-pages-plugin/maintain-contract/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/odere-pro/claude-wiki-pages-plugin/maintain-contract"><img src="https://agentmods.dev/badge/skills/odere-pro/claude-wiki-pages-plugin/maintain-contract.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 | $0.00118 | $0.01515 |
| Opus 5 | $0.00059 | $0.00758 |
| Sonnet 5 | $0.00024 | $0.00303 |
| Haiku 4.5 | $0.00012 | $0.00152 |
Grade A, and why
maintain-contract 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 9d 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 — 87 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Maintain contract — how to drive the wiki safely
This is the procedure any agent (this plugin's or a third party's) follows to
operate on a vault without corrupting it. It pairs with
/claude-wiki-pages:engine-api, which documents the tool surface this contract sequences.
The three invariants
- Ground, then judge, then verify. Compute facts with the engine (verify, fix, and the planned link-suggest/search) before reasoning. The LLM makes only judgment calls — topic placement, prose, what to synthesize — over engine-computed facts, never from memory. Close every write with
verify/heal. raw/is immutable. Never write, move, or delete anything undervault/raw/. Sources are the provenance anchor. Theprotect-rawhook enforces this; do not fight it.- Git is the safety net, not approval. Self-heal is automatic. Before structural changes, a checkpoint commit is written; rollback is
git revert <healCommit>. Do not prompt the user for permission to fix structure. Write phases that happen outside the engine (ingest writes, judgment fixes, polish) are bounded the same way:snapshot.sh prebefore the phase,snapshot.sh postafter it, and theSubagentStopcommit backstop sweeps up anything left dirty — no LLM write escapes git coverage.
Ingest (add knowledge)
snapshot.sh pre— checkpoint the pre-write state. 2. Read each source inraw/completely. 3. Write cited wiki pages (sources:as[[wikilinks]]to_sources/summaries), thensnapshot.sh post --label "ingest …". 4. Runengine.sh heal— it checkpoints, then verify→fix→re-verify, then commits. 5. Surface only what needs editorial intent (ambiguous merges, deletions).
Retrieve (answer questions)
- Use grounded retrieval (engine
search, orgrepoverwiki/untilsearchships) to fetch candidate pages. 2. Answer only from those pages. 3. Cite every claim with the[[Page Title]]it came from. 4. Never invent a citation; if the wiki cannot answer, say so.
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.
- 9d ago First seen · 87 lines · 118 tokens per session scan A b525d453c5de
maintain-contract is a skill published in the GitHub repository odere-pro/claude-wiki-pages-plugin (1 stars, last pushed 2mo ago), licensed MIT. It adds 118 tokens to every session and 1,515 once invoked, about $0.0006 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 2026-08-31.
Other skills, from other repositories
knowledge-garden
This skill should be used when the user asks to audit, health-check, or structurally validate one or more SPECIFIC named notes or a bounded topic cluster — 'audit these notes', 'check this note for orphans or broken links', 'fourth-wall check on [note]', 'validate the structure of [note]', 'spot-check [note]'. Runs a…
knowledge-maintain
This skill should be used when the user asks to fix, repair, or tidy one or more SPECIFIC named notes — 'fix these notes', 'fix the issues in [note]', 'add the missing relations to [note]', 'tidy up [note]', 'fix orphan [note]', 'apply the gardener findings for [note]'. Applies structural fixes (missing sections…
session-reflect
This skill should be used when the user asks to 'reflect on this session', 'save insights to memory', 'capture what we learned', 'commit this to memory', 'save decisions to Basic Memory', 'write up what we did', 'preserve what we found', 'save insights from this conversation', 'save what was discovered', 'capture…
knowledge-prime
This skill should be used when the user asks to 'prime context', 'load project knowledge', 'what do we know about this project', 'knowledge brief', 'project context', 'what packages are documented', 'show coverage for this project', 'dependency coverage report', 'which of our deps have notes', 'knowledge primer'…
nudge
This skill (explicit /nudge only) manages Claude Code feature-adoption nudges sourced from the Basic Memory note main/reference/claude-code-noteworthy-features. Bare /nudge = Mode A (sync): re-read the note, filter out features already marked adopted or declined, and regenerate the tip cache…
knowledge-ask
This skill should be used when the user asks 'what do I know about [topic]', 'what does Basic Memory say about [X]', 'do I have notes on [X]', 'look up [X] in my knowledge graph', 'ask memory about [X]', 'recall what we captured about [topic]', 'find notes on [topic]', 'knowledge question', 'query the knowledge…