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 AcKeskin/contexture --skill retrospect-coregit clone --depth 1 https://github.com/AcKeskin/contextureWrote 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/ackeskin/contexture/retrospect-core)<a href="https://agentmods.dev/skills/ackeskin/contexture/retrospect-core"><img src="https://agentmods.dev/badge/skills/ackeskin/contexture/retrospect-core.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 | $0.00084 | $0.02983 |
| Opus 5 | $0.00042 | $0.01491 |
| Sonnet 5 | $0.00017 | $0.00597 |
| Haiku 4.5 | $0.00008 | $0.00298 |
Grade A, and why
retrospect-core 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 — 152 lines — stays where its author put it; the contents beside it link to each section on GitHub.
retrospect-core
The meta-review engine. It is to checkpoint what deliver is to discover/prep/review: one mechanism, its consumers kept single-responsibility without duplicating the engine.
It does not know how to find decision drift or organ overlap — that domain logic lives in the calling skill's passes. retrospect-core owns the four things every meta-review run shares: orient, diff against the prior run, render per the review output contract, and route each confirmed finding to the organ that actually fixes it.
When to run
- A meta-review skill (
checkpoint) invokes one of the functions below programmatically. - Do not respond to user prose triggers. There is no
/retrospect-corecommand and no "when the user says X" branch. It is a format-and-orchestration helper. - Stateless per call, with one exception:
persistwrites a report artefact to disk (the substrate the next run'sdiffreads).
The contract
The caller (checkpoint) owns:
- Which corpus to point at and where its files live.
- The passes — the domain logic that produces findings.
retrospect-core owns:
orient(target)— inventory + since-last-run delta.diff(findings, baselinePath)— NEW/CARRIED/RESOLVED/WONT_FIX tagging.render(report)— the review-output-contract report body.route(finding)— propose-confirm-commit, then hand off to capture / memory-audit / a proposals stub.persist(report, scope)— writelatest.md+v<N>.mdfor the next run's baseline.
A finding passed between caller and engine has this shape:
Finding = {
id: "R<NNN>", // assigned by diff(); provisional in scan order before that
pass: string, // caller's pass label — e.g. "decision-integrity", "responsibility-overlap"
locator: string, // the thing the finding is about: a proposal id, memory path, skill name, file:line
verdict: string, // pass-specific tag — e.g. HOLDS / SUPERSEDED-unmarked / DEAD / OVERLAP
severity: "Critical"|"High"|"Medium"|"Low",
what: string, // one-line, specific
route: "capture"|"memory-audit"|"proposal"|"direct-fix"|"none",
proposed_action: string, // concrete — the capture content, the back-link edit, the proposal stub title
status?: string, // open | applied | skipped | wont_fix:<reason> (set during route())
}
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 · 152 lines · 84 tokens per session scan A 6ac2995f11cc
retrospect-core is a skill published in the GitHub repository AcKeskin/contexture (2 stars, last pushed 1mo ago), licensed MIT. It adds 84 tokens to every session and 2,983 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 2026-09-03.
Other skills, from other repositories
ocr
AI-powered multi-agent code review. Simulates a team of Principal Engineers reviewing code from different perspectives. Use when asked to review code, check a PR, analyze changes, or perform code review.
generic-react-code-reviewer
Review React/TypeScript code for bugs, security vulnerabilities, performance issues, accessibility gaps, and CLAUDE.md workflow compliance. Enforces TypeScript strict mode, GPU-accelerated animations, WCAG AA accessibility, bundle size limits, and surgical simplicity. Use when completing features, before commits, or…
generic-fullstack-code-reviewer
Review full-stack code for bugs, security vulnerabilities, performance issues, accessibility gaps, and CLAUDE.md compliance. Enforces TypeScript strict mode, input validation, GPU-accelerated animations, and design system consistency. Use when completing features, before commits, or reviewing pull requests.
generic-static-code-reviewer
Review static site code for bugs, security issues, performance problems, accessibility gaps, and CLAUDE.md compliance. Enforces pure HTML/CSS/JS standards, minimal page weight, mobile-first design. Use when completing features, before commits, or reviewing changes.
OCR Review-to-Approval Loop
Drive a PR to an approved code review by looping OCR's multi-agent review and address steps. Runs /ocr:review then /ocr:address repeatedly until the review verdict is APPROVE, then one final /ocr:address for leftover suggestions, posting every review and every address round to the GitHub PR as comments. Use when the…
code-reviewer
Review code for best practices, bugs, and security issues.