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/ackeskin/contexture/prepgit clone --depth 1 https://github.com/AcKeskin/contextureWhat 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 | $0.00024 | $0.00293 |
| Opus 5 | $0.00012 | $0.00147 |
| Sonnet 5 | $0.00005 | $0.00059 |
| Haiku 4.5 | $0.00002 | $0.00029 |
Grade A, and why
prep 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.
What it actually says
Run the prep skill for the current task.
Any text after /prep is treated as scope hints that bias scope detection. Examples:
/prep— infer scope from the current task + project./prep typescript ui-components— prime for TypeScript UI work explicitly./prep cpp rendering— prime for C++ rendering work, overriding cwd-based inference.
The skill identifies task scope (language, domain, project, task type), calls skills/discover/SKILL.md with kind: "architectural-rule" and render_bodies: true, merges and caps the result (project > domain > language > universal; ≤ 20 rules; <500 tokens), reads .claude/architecture.md if present, and surfaces the priming block with the loaded rules so you can confirm the scope is right.
Prep also auto-fires on first substantive task of a session, first task after /clear, and when the user signals a topic shift. During subsequent work it observes drift and asks before proceeding outside the primed scope — never silently. On user push-back ("you violated rule X"), prep proposes a capture via /capture rather than writing directly.
See ~/.claude/skills/prep/SKILL.md for the full procedure.
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 · 18 lines · 24 tokens per session scan A c13df3006845
prep is a command published in the GitHub repository AcKeskin/contexture (2 stars, last pushed 1mo ago), licensed MIT. It adds 24 tokens to every session and 293 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 2026-08-31.
Other commands, from other repositories
translate-review-to-single-human
Translate a multi-reviewer code review into a single human-voice GitHub PR comment.
address
Address code review feedback — corroborate, validate, and implement changes from a review's final.md.
map
Generate a Code Review Map to help navigate large, complex changesets.
review
Run an AI-powered multi-agent code review on your changes.
sync-reviewers
Sync reviewer metadata from markdown files to reviewers-meta.json for the dashboard.
post
Post the current OCR review to a GitHub PR.