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 agentmods add agents/stahl-g/briefloop/harness-subagentsgit clone --depth 1 https://github.com/Stahl-G/briefloopWrote 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/stahl-g/briefloop/harness-subagents)<a href="https://agentmods.dev/agents/stahl-g/briefloop/harness-subagents"><img src="https://agentmods.dev/badge/agents/stahl-g/briefloop/harness-subagents.svg" alt="Measured on agentmods" 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 | $0.00000 | $0.00356 |
| Opus 5 | $0.00000 | $0.00178 |
| Sonnet 5 | $0.00000 | $0.00071 |
| Haiku 4.5 | $0.00000 | $0.00036 |
Grade A, and why
harness-subagents 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 3d 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.
What it actually says
Harness Subagents
This project has two harness layers.
1. Draft Audit Harness
Purpose:
auditor subagent -> CompositeAuditAgent -> DeterministicAuditAgent -> QualityHarnessAuditAgent -> optional semantic audit adapter
Responsible subagent:
draft-audit-harness
It reviews and implements:
- missing or orphan claim IDs
- number/source coverage
- reporting-window freshness
- missing source dates
- placeholders
- internal workflow residue
- process residue
- metadata residue
- unsupported certainty wording
- investment-advice style language
- needs_recrawl leakage
- T5 or low-confidence source leakage
- low numeric source density
- possible unit inflation
- repeat/background claims in executive summary
- stale/no-update filler
2. Final Delivery Harness
Purpose:
draft facts pass -> final text quality pass -> rendered document quality pass -> delivery allowed
Responsible subagents:
final-quality-harness
rendered-output-harness
final-quality-harness checks:
- final report depth
- quiet_week handling
- front-page metadata
- stale-current framing
- executive summary bullet separation
- internal workflow residue
- publication readiness
rendered-output-harness checks:
- DOCX/PDF rendering depth
- heading mapping
- bullet separation after rendering
- footer/page fields
- margins
- wide table conversion
- missing rendering dependencies
Rule
Correct facts are necessary but not sufficient for final delivery. A report can pass source audit and still be blocked by final-quality or rendered-output gates.
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.
- 3d ago First seen · 78 lines · 0 tokens per session scan A df88af7f5d39
harness-subagents is an agent published in the GitHub repository Stahl-G/briefloop (9 stars, last pushed 5d ago), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 356 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 2026-08-31.
Other agents, from other repositories
hormozi-orchestrator
Master offer-building orchestrator inspired by Alex Hormozi's frameworks. Captures raw ideas, notes, or existing offers from the user, interviews them to extract market, problem, outcome, and constraints, then routes to specialized subagents to produce offer documents. Use when user wants to build an offer, validate a…
pdf-debugger
Debug PDF processing issues — OCR failures, text extraction problems, AI misclassification, and renaming errors. Use when a PDF isn't being renamed correctly.
technical-design-specialist-subagent
Authors Technical Design Documents — converts SRS/feature specs into architecture, data models, API contracts, ADRs with CodeGraph impact analysis. Triggers: technical design, architecture/system design, design spec.
startup-founder-primary-agent
Startup founder operations — reports, quotations, spreadsheets, presentations, day-to-day business.
startup-ceo-subagent
Startup-style PowerPoint presentations — pitch decks, investor slides, board updates.
stage-gate-controller
Loop Craft 的质量、工程与过程管控 Agent。在每个 stage gate 前后独立复核方向是否偏移、产出是否名副其实、预算是否失控。必须在任何一个模块被宣布完成、任何一次 commit 被提议、任何一次范围变更被采纳之前调用。它只裁决，不实现。.