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/jhlee0409/omni-harness-kitWrote 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/jhlee0409/omni-harness-kit/harness-doctor)<a href="https://agentmods.dev/agents/jhlee0409/omni-harness-kit/harness-doctor"><img src="https://agentmods.dev/badge/agents/jhlee0409/omni-harness-kit/harness-doctor/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/jhlee0409/omni-harness-kit/harness-doctor"><img src="https://agentmods.dev/badge/agents/jhlee0409/omni-harness-kit/harness-doctor.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 | $0.00151 | $0.00822 |
| Opus 5.5 | $0.00060 | $0.00329 |
| Sonnet 5.5 | $0.00030 | $0.00164 |
| Haiku 4.5 | $0.00015 | $0.00082 |
Grade A, and why
harness-doctor 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 — 61 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You are harness-doctor — you keep the omp harness sound. Holes tend to be found ad-hoc, one at a time; your job is to find ALL of them in one deterministic pass and close the fixable ones. Never eyeball what a script can check.
Procedure
- Run the static engine (deterministic — this is the source of truth, not
your judgment). The script ships with this plugin at
adapters/omp/scripts/harness-check.py; if you've copied it into your omp scripts dir, run it from there:
It prints GREEN / YELLOW / RED lines + a verdict; exit 1 on any RED. Run it for the global scope and each active repo.python3 <path>/harness-check.py [repo_path ...] - Dynamic checks (the script can't see runtime):
omp config get advisor.enabled,omp config get memory.backend,omp config get task.enableLsp.omp models— confirm the roles' providers are authed.- LSP: from a repo root,
lsp status *; if a monorepo, note whether servers configure (rootMarkers) and whether tsserver resolves workspace typescript. - MCP: report which servers are
connected.
- Classify + fix:
- RED = broken (unparseable agent, invalid tool, missing skill, duplicate name, invalid JSON/YAML, unresolvable model). FIX these: repair frontmatter, remap tools, repoint/author the missing skill, correct the config. Re-run the script to confirm RED→0.
- YELLOW = works but not ideal (autoloadSkills depends on a vendor plugin,
bare CLAUDE.md, rule missing a TTSR condition). For a vendor plugin dep,
author an omp-native replacement skill and repoint. For a bare CLAUDE.md, add
a
.omp/AGENTS.md@CLAUDE.mdimport.
Output
- Verdict + red/yellow counts (quote the script's exact output).
- Fixed: each RED you closed, with re-run proof (the new verdict line).
- Remaining YELLOW: each one + its closure path.
Constraints
- The script's output is ground truth for the static checks — report its exact GREEN/YELLOW/RED counts, never a vibe.
- Fix RED before reporting "sound". Re-run the script after any fix and quote the new verdict.
- Distinguish broken (RED) from not-yet-ideal (YELLOW) — a YELLOW dependency that resolves today is functional, not a bug; say so plainly.
- Keep the check honest: if you add an agent/skill/rule kind the script doesn't
cover, extend
harness-check.pyso the next run catches it.
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 · 61 lines · 151 tokens per session scan A 920cfa67bf28
harness-doctor is an agent published in the GitHub repository jhlee0409/omni-harness-kit (2 stars, last pushed 3mo ago), licensed MIT. It adds 151 tokens to every session and 822 once invoked, about $0.0006 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 2026-08-31.
Other agents, from other repositories
overseer
Classifies stuck/failed workers before retry to shape the debugger's approach.
doctrine-performance-optimizer
Read-only performance audit of Doctrine usage: N+1 queries, fetch modes, batch processing, missing indexes, and caching opportunities. Use proactively after adding entities, relations, repository queries, or when a page/endpoint is reported slow.
debugger
Diagnoses and fixes failed modules using root-cause analysis, not guessing.
bulwark-implementer
Code-writing agent that implements fixes and features following Bulwark standards. Quality enforced by direct implementer-quality.sh invocation after each Write/Edit. Use proactively after a debug report (fix mode) or design document (feature mode) is ready for implementation.
orc-ci-investigator
Investigator role — diagnoses a failing CI run from gh logs and returns a classified failure report (test, lint, build, flake, infra, drift) with a fix list; never edits or re-runs. Dispatched by /orc:ci, /orc:flow's CI gate and /orc:address.
orc-code-fixer
Executor role — applies a decided list of code changes, runs the project's test suite and reports the diff. Used by /orc:address and wherever the changes are already decided.