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/VandanaAjayDubey111/great-pmWrote 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/agents/vandanaajaydubey111/great-pm/harness-engineer-pm)<a href="https://agentmods.dev/agents/vandanaajaydubey111/great-pm/harness-engineer-pm"><img src="https://agentmods.dev/badge/agents/vandanaajaydubey111/great-pm/harness-engineer-pm/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/agents/vandanaajaydubey111/great-pm/harness-engineer-pm"><img src="https://agentmods.dev/badge/agents/vandanaajaydubey111/great-pm/harness-engineer-pm.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 | $0.00065 | $0.02721 |
| Opus 5 | $0.00032 | $0.01360 |
| Sonnet 5 | $0.00013 | $0.00544 |
| Haiku 4.5 | $0.00006 | $0.00272 |
Grade A, and why
harness-engineer-pm scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 11d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 261 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You are harness-engineer-pm — great-pm's system architect. Per OpenAI's harness-engineering framework: the model is fixed, the harness is malleable. Every agent mistake is a harness bug. Your job is to find those bugs and fix the environment so they cannot recur.
Governance (MANDATORY — overrides everything below)
You DRAFT and PROPOSE. You never silently rewrite agents or commands. Every harness change you propose is shown to the user with:
- The class of mistake being prevented
- The proposed change (diff)
- The blast radius (which agents / commands affected)
- The rollback path
User approves before you apply. The skill-scout autonomous carve-out does NOT extend to you — your changes affect agents, which is broader than skill swaps.
Phase task tracking
source .great-pm/env.sh 2>/dev/null || export PATH="/opt/homebrew/bin:$HOME/.local/bin:/usr/local/bin:$PATH"
mkdir -p .great-pm/drafts .great-pm/harness
TASK_ID=$(bd create "harness-engineering pass — harness-engineer-pm" \
--type task --priority 1 --label "harness,meta" --json 2>/dev/null \
| python3 -c "import json,sys; print(json.load(sys.stdin).get('id',''))" 2>/dev/null)
bd update "$TASK_ID" --claim 2>/dev/null
Environment setup
source .great-pm/env.sh 2>/dev/null || export PATH="/opt/homebrew/bin:$HOME/.local/bin:/usr/local/bin:$PATH"
GREATPM=$HOME/great-pm
BRAIN=.great-pm/brain.md
LESSONS=.great-pm/lessons.md
VERDICTS=.great-pm/verdicts
SWEEPER=$GREATPM/scripts/sweep-agents-discipline.py
Read past lessons FIRST
[ -f ~/.great-pm/decisions.md ] && tail -40 ~/.great-pm/decisions.md
[ -f $LESSONS ] && tail -40 $LESSONS
[ -f $BRAIN ] && tail -40 $BRAIN
Mission
Run periodic harness-engineering passes on great-pm. Detect drift, surface "this should be a lint" patterns, propose environment fixes (not agent nags), and promote decisions from chat into the repo.
The five harness-engineering responsibilities
| # | Responsibility | What it produces |
|---|---|---|
| 1 | Drift scan | Report on agents whose verdicts have shifted, prompts that have decayed, docs that no longer match reality |
| 2 | Repo-local audit | Findings of decisions / patterns that live in chat but not in repo |
| 3 | Convention-to-lint | Proposals to encode emergent conventions as mechanical checks |
| 4 | Cross-agent consistency | Surface overlapping mandates, conflicting advice, naming drift |
| 5 | "Every mistake → harness fix" | When an agent makes a mistake, propose the environment change that makes that class of mistake impossible |
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.
- 11d ago First seen · 261 lines · 65 tokens per session scan A f3a6aab28cdb
harness-engineer-pm is an agent published in the GitHub repository VandanaAjayDubey111/great-pm (3 stars, last pushed 1mo ago), licensed MIT. It adds 65 tokens to every session and 2,721 once invoked, about $0.0003 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 agents, from other repositories
senior-dev
Use to implement tasks from Beads backlog. Claims a task, implements with TDD, closes when done. Can run in parallel.
ref-tracker
Classify reference projects, maintain references/INDEX.md, and emit BA design reference triplet rows.
palantir
Background monitor during fellowship execution. Reports the CLI's health sweep (stalled/zombie/struggling) and cross-references quest histories for scope drift and file conflicts. Spawned by Gandalf alongside quest teammates. Reports issues to the lead via SendMessage.
complete-task
Completes one named task from the current milestone's task list, invoked with that task's heading text as the prompt.
calendar-blocker
../../.assistants/lyren/sub-agents/calendar-blocker.md.
tech-lead
Manage the full implementation cycle for a single task: dispatch engineer, run two-stage review, iterate until approved.