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 grcengineering/companion --skill task-retrospectivegit clone --depth 1 https://github.com/grcengineering/companionWrote 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/grcengineering/companion/task-retrospective)<a href="https://agentmods.dev/skills/grcengineering/companion/task-retrospective"><img src="https://agentmods.dev/badge/skills/grcengineering/companion/task-retrospective/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/grcengineering/companion/task-retrospective"><img src="https://agentmods.dev/badge/skills/grcengineering/companion/task-retrospective.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 | $0.00066 | $0.00634 |
| Opus 5.5 | $0.00026 | $0.00254 |
| Sonnet 5.5 | $0.00013 | $0.00127 |
| Haiku 4.5 | $0.00007 | $0.00063 |
Grade A, and why
task-retrospective 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 1mo 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 — 65 lines — stays where its author put it; the contents beside it link to each section on GitHub.
task-retrospective
What
Turn work the learner already did into structured learning. The work can be real local material in a local adapter, but the output is reflection and reusable learning, not operational judgement.
When
- The learner describes a completed GRC task.
- The learner points at local output from a completed task.
- The learner wants to learn from a painful or slow workflow.
- The learner asks what to do better next time after they already acted.
Not For
- Live vendor review, audit prep, control testing, policy authoring, or programme decisions.
- Fictional case practice before the learner has acted. Use
practice-scenario. - General stuckness without a completed task. Use
socratic-coach.
Inputs
- Learner's description of completed work.
- Real local artefacts only when intentionally provided in a local adapter.
brain/task-extraction.md.- Optional GRC primitives or corpus files for citation grounding.
Steps
- Confirm that this is a retrospective, not operational advice.
- Use real local output as learning material when intentionally provided.
- Ask the six extraction questions from
brain/task-extraction.md, one or two at a time. - Surface two or three patterns in the learner's process.
- Connect the pattern to GRC Engineering primitives or corpus ideas where relevant.
- Produce a small reusable learning artefact or proposed local profile/progress update.
- End with: "Where else would this pattern apply?"
Validation
- The output assesses the learner's process, not the actual outcome.
- Sensitive details are absent or kept local and abstracted.
- The learner leaves with a reusable note, checklist, pattern map, or progress/profile proposal.
Gotchas
- A real local artefact is learning material, not permission to operate the programme.
- If the learner asks "was this vendor acceptable?", refuse that decision and reflect on the reasoning process.
- Do not ask for all six extraction answers at once unless the learner requests a worksheet.
What ships with it
3 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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.
- 1mo ago First seen · 65 lines · 66 tokens per session scan A 611a2df5bda2
task-retrospective is a skill published in the GitHub repository grcengineering/companion (35 stars, last pushed 5mo ago), licensed MIT. It adds 66 tokens to every session and 634 once invoked, about $0.0003 per session on Opus 5.5. 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 2026-08-30.
Other skills, from other repositories
control-mapping
Maps NIST controls to FedRAMP requirements and documents. Use when helping with control implementation, compliance mapping, security baseline alignment, or understanding control requirements.
frmr-analysis
Analyzes FedRAMP FRMR documents to extract control mappings, KSI entries, and version changes. Use when the user asks about FedRAMP requirements, control mappings, compliance data, or needs to understand FRMR document content.
performing-soc2-type2-audit-preparation
Automates SOC 2 Type II audit preparation including gap assessment against AICPA Trust Services Criteria (CC1-CC9), evidence collection from cloud providers and identity systems, control testing validation, remediation tracking, and continuous compliance monitoring. Covers all five TSC categories (Security…
bar-prep-questions-bossmann007
Questões de preparação para o Exame de Ordem (OAB) — 1ª fase (objetiva) ou 2ª fase (peça/discursiva), com foco nas matérias fracas do aluno. Rastreia erros e volta aos padrões. Use quando o usuário disser "prova da OAB", "questões de 1ª fase", "peça prática", ou "me teste pra Ordem".
irac-practice-bossmann007
Corrige uma peça/parecer no método IRAC (ou estrutura equivalente) quanto a estrutura, identificação de teses, precisão de fundamentação, profundidade de análise e organização. NÃO reescreve o texto nem mostra resposta-modelo; rastreia padrões entre sessões. Use quando o usuário disser "corrige meu IRAC", "avalia meu…
study-plan
Build or update a long-term bar prep (or exam prep) study plan — phases, subjects weighted by weakness, daily session schedule, adaptive to session history in study-plan.yaml. Use when the user says "build a study plan", "plan my bar prep", "schedule my studying", or "how should I study for [X]".