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 juliusz-cwiakalski/agentic-delivery-os --skill review-decisiongit clone --depth 1 https://github.com/juliusz-cwiakalski/agentic-delivery-osWrote 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/juliusz-cwiakalski/agentic-delivery-os/review-decision)<a href="https://agentmods.dev/skills/juliusz-cwiakalski/agentic-delivery-os/review-decision"><img src="https://agentmods.dev/badge/skills/juliusz-cwiakalski/agentic-delivery-os/review-decision/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/juliusz-cwiakalski/agentic-delivery-os/review-decision"><img src="https://agentmods.dev/badge/skills/juliusz-cwiakalski/agentic-delivery-os/review-decision.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 | $0.00033 | $0.01220 |
| Opus 5 | $0.00016 | $0.00610 |
| Sonnet 5 | $0.00007 | $0.00244 |
| Haiku 4.5 | $0.00003 | $0.00122 |
Grade A, and why
review-decision 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 12d 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 — 83 lines — stays where its author put it; the contents beside it link to each section on GitHub.
User invocation:
/review-decision
where <ID> is the decision record identifier (e.g., ADR-0001) or the record's number for the relevant type.
This command is read-only by default: it produces a review artifact/verdict in its response and does NOT write, edit, stage, or commit any file. The caller decides whether and how to act on the findings (e.g., feed them back to @decision-advisor via /write-decision update, or escalate to a human reviewer for R3).
<record_resolution>
- Parse the ID into TYPE + zeroPad4 (e.g.,
ADR-0001→ typeADR, number0001). If only digits are given, default the type to ADR only when the type is genuinely unspecified. - Locate the record file by scanning
doc/decisions/<TYPE>-<number>-*.md. - If not found, return
NOT_FOUNDwith guidance to run/plan-decision+/write-decisionfirst, and stop. </record_resolution>
<independence_note>
State at the top of the review which independence level applies (per @decision-critic's honesty framing):
- Single-model configuration: the critic is a first-pass check, NOT independent assurance (same-model/same-prompt lineage is not independent evidence).
- Multi-model configuration: the critic runs on a different model family, providing genuine independence (recommended for R3, not mandated).
Regardless of the verdict, R3 decisions ALWAYS require a human reviewer. A PASS from the critic does not authorize an R3 decision. </independence_note>
<output_contract> Return a structured review artifact (in the response only — no file writes):
- Status:
REVIEWED|NOT_FOUND|NEEDS_INPUT - Record:
<TYPE>-<number>and path reviewed - Verdict: PASS | PASS_WITH_RISKS | REWORK
- Independence note: single-model (first-pass check) | multi-model (independent) — plus the R3-requires-human-review reminder
- Findings: framing / option-space / constraint-compliance / assumption-sensitivity / premortem / stakeholder-harm / confidence / automation-bias
- Accepted risks (for PASS_WITH_RISKS): explicit list the human decider must accept
- Required changes (for REWORK): explicit, actionable list
- Next step: e.g., "Return findings to @decision-advisor to address" / "Escalate to human reviewer (R3)" / "Decision is sound; proceed to human authorization" </output_contract>
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.
- 12d ago First seen · 83 lines · 33 tokens per session scan A 609e74b34521
review-decision is a skill published in the GitHub repository juliusz-cwiakalski/agentic-delivery-os (38 stars, last pushed 2d ago), licensed MIT. It adds 33 tokens to every session and 1,220 once invoked, about $0.0002 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 2026-08-30.
Other skills, from other repositories
Verification & Quality Assurance
Comprehensive truth scoring, code quality verification, and automatic rollback system with 0.95 accuracy threshold for ensuring high-quality agent outputs and codebase reliability.
agent-github-modes
Agent skill for github-modes - invoke with $agent-github-modes.
workflow-automation
Workflow creation, execution, and template management. Automates complex multi-step processes with agent coordination. Use when: automating processes, creating reusable workflows, orchestrating multi-step tasks. Skip when: simple single-step tasks, ad-hoc operations.
deployment-pipeline-design
Design multi-stage CI/CD pipelines with approval gates, security checks, and deployment orchestration. Use this skill when designing zero-downtime deployment pipelines, implementing canary rollout strategies, setting up multi-environment promotion workflows, or debugging failed deployment gates in CI/CD.
gitlab-ci-patterns
Build GitLab CI/CD pipelines with multi-stage workflows, caching, and distributed runners for scalable automation. Use when implementing GitLab CI/CD, optimizing pipeline performance, or setting up automated testing and deployment.
bazel-build-optimization
Optimize Bazel builds for large-scale monorepos. Use when configuring Bazel, implementing remote execution, or optimizing build performance for enterprise codebases.