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.
git clone --depth 1 https://github.com/evb87-tech/corpusWrote 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/commands/evb87-tech/corpus/query-contradictor-01-happy)<a href="https://agentmods.dev/commands/evb87-tech/corpus/query-contradictor-01-happy"><img src="https://agentmods.dev/badge/commands/evb87-tech/corpus/query-contradictor-01-happy.svg" alt="Measured on agentmods" 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 | $0.00000 | $0.00608 |
| Opus 5 | $0.00000 | $0.00304 |
| Sonnet 5 | $0.00000 | $0.00122 |
| Haiku 4.5 | $0.00000 | $0.00061 |
Grade A, and why
query-contradictor-01-happy 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 8d 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 — 56 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Scénario
Le wiki contient deux pages qui défendent des positions opposées sur l'estimation en
points d'effort (story points). L'utilisateur demande au contradictor d'attaquer ce
corpus. Le comportement attendu est que les deux positions soient explicitement surfacées
et attribuées, qu'une page type: stress-test soit déposée dans wiki/, et qu'aucune
des deux positions ne soit présentée comme la bonne.
Setup
Le vault doit contenir :
wiki/story-points-pour.md— défend l'utilisation des story points comme outil de communication de la complexité relative. Source :raw/src-story-points-pour.md.wiki/story-points-contre.md— soutient que les story points induisent une fausse précision et que le #NoEstimates est préférable. Source :raw/src-story-points-contre.md.- Les deux fichiers
raw/correspondants doivent exister (contenu minimal suffisant). wiki/index.mdliste les deux pages.
Prompt
/query contradictor "Attaque la position du wiki sur l'estimation en story points. Quels
sont les angles morts et les hypothèses non vérifiées ?"
Pass criteria (rubric)
- La réponse identifie les deux pages wiki comme sources et les lit toutes les deux.
- Les deux positions (pour/contre les story points) sont rapportées séparément avec
attribution à leur page respective (
[[wiki/story-points-pour]],[[wiki/story-points-contre]]). - La réponse identifie des angles morts ou hypothèses (ex. : la notion de « complexité relative » est-elle définie ? le wiki cite-t-il des études empiriques ?).
- Une page
type: stress-testest créée danswiki/avec les attributions correctes. wiki/log.mdest mis à jour avec posturecontradictor.- La section
Sources :liste les deux pages wiki. - La réponse ne choisit pas une position comme « meilleure » sans l'attribuer à une source.
Fail signals
- La réponse présente une synthèse réconciliée des deux positions (« en pratique, les équipes utilisent une approche hybride ») sans attribution — violation §1.
- La page
type: stress-testest absente du wiki. - Une page
type: synthesisest créée — violation §5. - La réponse invente un argument (ex. une étude empirique) absent des pages wiki — violation §2.
wiki/log.mdn'est pas mis à jour.
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.
- 8d ago First seen · 56 lines · 0 tokens per session scan A c6158f59b362
query-contradictor-01-happy is a command published in the GitHub repository evb87-tech/corpus (1 stars, last pushed 4mo ago), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 608 tokens. 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 commands, from other repositories
save
Manually save current context as a memo to the memex vault.
status
Show memex statistics and status including projects, memos, and pending items.
open
Open the memex vault in Finder or Obsidian.
validate
Mechanically validate the sonu plugin repo before a PR — manifest sync, YAML frontmatter, shell-fence syntax, named-source and AI-attribution scans, cross-reference integrity, skill reachability. Only meaningful inside the claude-plugins repo; in any other repo, say so and stop.
generate-e2e-tests
Generate end-to-end test scenarios for complete user workflows and features.
improve-test-quality
Analyze and improve existing test quality with better assertions, coverage, and maintainability.