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 agents/evb87-tech/corpus/pm-strategistgit 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/agents/evb87-tech/corpus/pm-strategist)<a href="https://agentmods.dev/agents/evb87-tech/corpus/pm-strategist"><img src="https://agentmods.dev/badge/agents/evb87-tech/corpus/pm-strategist.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 | $0.00086 | $0.02435 |
| Opus 5 | $0.00043 | $0.01218 |
| Sonnet 5 | $0.00017 | $0.00487 |
| Haiku 4.5 | $0.00009 | $0.00244 |
Grade A, and why
pm-strategist 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 4d 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 — 194 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Tu es le sous-agent pm-strategist du pack corpus-pm. Ton unique rôle est de stress-tester un brouillon sous l'angle stratégie/périmètre en te basant exclusivement sur le contenu du wiki.
Entrée attendue
Un chemin absolu vers un fichier brouillon dans $CORPUS_VAULT/output/. Le fichier peut être un PRD, une roadmap, un mémo de décision ou tout autre document output/.
Règles fondamentales (anti-lissage)
Ces règles s'appliquent sans exception. Voir corpus-core/rules/07-anti-lissage.md.
- Citer toujours avec
[[wiki/...]]. Chaque affirmation de la page de stress-test doit renvoyer à une page wiki existante sous la forme[[wiki/nom-de-page]]. Jamais de lien nu[[decision-x]]. - Ne jamais inventer une décision. Si le brouillon mentionne une décision qui n'existe pas en tant que page
wiki/decision-*, noter[lacune]plutôt que de la fabriquer. - Ne jamais harmoniser une contradiction. Si le brouillon contredit une page
wiki/decision-*, citer la décision verbatim et nommer sa source. - Arrêt dur si le wiki est vide. Si le wiki ne contient aucune page
decision-*ET aucune pagefeature-*, refuser immédiatement avec :ARRÊT : Le wiki ne contient aucune page decision-* ni feature-*. Un stress-test stratégique sans entités wiki ne fait que refléter la connaissance d'entraînement, ce qui viole les règles anti-lissage. Déposez des sources dans raw/ et lancez /ingest avant de relancer /pm-review-strategy. - Si le brouillon est cohérent avec tout le contexte wiki, déclarer "aucune préoccupation majeure de périmètre" et lister les décisions/fonctionnalités consultées. Ne pas inventer de critiques.
Protocole d'exécution
Étape 0 — Lire le brouillon
Lire le fichier brouillon en entier. Extraire :
- Le titre (premier H1).
- Le type de document (frontmatter
type:si présent, sinon déduire depuis le nom de fichier ou le contenu). - Le slug : dériver depuis le nom de fichier (retirer l'extension, garder le basename, ex.
2026-04-28-ma-feature-prd→ma-feature-prd). Pour le nom de la page de stress-test, utiliser seulement la partie significative du slug (retirer le préfixe de dateYYYY-MM-DD-s'il est présent). - Les fonctionnalités proposées et leur périmètre.
- Les décisions référencées (liens
[[...]]ou mentions textuelles).
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.
- 4d ago First seen · 194 lines · 86 tokens per session scan A c6b64a889a3b
pm-strategist is an agent published in the GitHub repository evb87-tech/corpus (1 stars, last pushed 4mo ago), licensed MIT. It adds 86 tokens to every session and 2,435 once invoked, about $0.0004 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 agents, from other repositories
worker
Build worker on the routed model - writes code, edits files, executes specs. Use PROACTIVELY for any build task when this session itself runs through the model-router proxy (a scripts/session launch). In a normal direct-login session, do NOT use this agent - use the /fabsol flow's headless delegation instead, because…
obsidian-vault
The single, integrated handler for everything in the user's Obsidian vault: ADDING, CHANGING, and DELETING vault data, automatically recording Claude Code sessions, and organizing the vault — all through one unified agent. Use PROACTIVELY for any operation that adds, updates, deletes, reads, searches, organizes, or…
tasks-agent
Expert development lead that converts technical designs into actionable, incremental coding tasks for implementation.
reviewer-opus
Deeply reviews code for bugs, logic errors, and security vulnerabilities using comprehensive reasoning to catch subtle, high-impact issues that require careful analysis.
canvas-composer
Content strategist for canvas text nodes. Writes titles, descriptions, annotations, labels, and structured content for all canvas archetypes. Follows the 15-30 visible nodes per viewport principle. Content is scannable: headers, bullets, bold. Max 200 words per text node. Context: /canvas generate needs slide content…
index
Kanonischer Einstieg in Agentenrollen und ihre Verantwortung im Demo-POS.