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.
git clone --depth 1 https://github.com/Abhillashjadhav/AI-PM-essential-skillsWrote 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/commands/abhillashjadhav/ai-pm-essential-skills/004-harden-pm-verifier-production-harness)<a href="https://agentmods.dev/commands/abhillashjadhav/ai-pm-essential-skills/004-harden-pm-verifier-production-harness"><img src="https://agentmods.dev/badge/commands/abhillashjadhav/ai-pm-essential-skills/004-harden-pm-verifier-production-harness/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/commands/abhillashjadhav/ai-pm-essential-skills/004-harden-pm-verifier-production-harness"><img src="https://agentmods.dev/badge/commands/abhillashjadhav/ai-pm-essential-skills/004-harden-pm-verifier-production-harness.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 | $0.00000 | $0.00543 |
| Opus 5.5 | $0.00000 | $0.00217 |
| Sonnet 5.5 | $0.00000 | $0.00109 |
| Haiku 4.5 | $0.00000 | $0.00054 |
Grade A, and why
004-harden-pm-verifier-production-harness 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 — 58 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Harden pm-verifier into an installable production CI harness
Outcome
Make pm-verifier the one installable, versioned evaluation product in this
repository. A clean checkout must support this complete loop without a second
evaluation product:
define suite -> execute isolated trials -> grade evidence -> inspect failures -> release decision
Verified starting point
PR #39 implements an evidence grader and release-decision engine. It validates pre-recorded trials, applies deterministic and calibrated model judgments, reports failure slices, and fails closed when evidence is invalid. It does not yet execute a system under test, install as a package, or expose transcript inspection through its CLI.
Required changes
- Add a dependency-free Python package with a
pm-verifierconsole script and semantic version output. - Add versioned suite/result schemas and explicit compatibility blocking.
- Add
inspectso a PM can read failing trials, trajectories, outcomes, grader reasons, slices, and clusters. - Replace the combined calibration ratio with separate binary and ordinal statistics. Require a minimum human-golden sample, chance-corrected label agreement, and an agreement confidence lower bound.
- Make trajectory and other deterministic checks explicitly gating or diagnostic. Exact paths must not gate by accident, while policy/safety path checks may remain deliberate release gates.
- Preserve binary risk gates and add partial quality scores for capability hill-climbing.
- Add a JSON-over-stdio subprocess adapter. Start a fresh process for every trial, never disclose expected answers to the adapter, require unique isolation IDs, and record an environment fingerprint.
- Add known-answer tests for degenerate judges, insufficient calibration, non-gating trace checks, adapter failures, trial isolation, schema mismatch, deterministic outputs, and secret redaction.
- Audit every useful public capability in
pm-evalsandEvals-pass-1before recommending repository removal.
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 · 58 lines · 0 tokens per session scan A 6777b8b45739
004-harden-pm-verifier-production-harness is a command published in the GitHub repository Abhillashjadhav/AI-PM-essential-skills (9 stars, last pushed 10d ago), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 543 tokens. 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 2026-09-16.
Other commands, from other repositories
judge-ci-red
Load when: loom-daemon forge wait-checks printed LOOM-CHECKS-RED, or post-verdict.sh … approved exited 6, and a CI-failing verdict is next.
write-test-plan
Generate or update change test plan.
test-setup
Scaffold test framework and CI pipeline.
ci-setup
Setup CI pipeline from scratch. Creates GitHub Actions or GitLab CI configuration with lint, test, build, and deploy stages.
harness:cicd-fleet
Autonomous CI/CD-remediation orchestrator — triage the red CI/CD-run and flaky-test backlog by cause, fan out worktree-isolated subagents that run the real deflake/heal pipeline, independently verify each fix by artifact and deterministic all-OS CI, and hand back a batch of remediation PRs for one bulk human review.…
check-fix
Execute quality gates, fix any issues found, and create a single high-quality Conventional Commit summarizing all changes made.