Better Harness is an open-source platform for studying and improving how coding agents work on tasks, using written harnesses, controlled tests, and evidence from projects and sessions. Coding teams use it to identify workflow problems and turn them into prioritized improvements and verifiable next steps. Its catalogue entries extend coding-agent workflows for supported hosts such as Claude Code, Codex, Cursor, and others.
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 QoderAI/better-harness --skill generate-harness-dslgit clone --depth 1 https://github.com/QoderAI/better-harnessWrote 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/qoderai/better-harness/generate-harness-dsl)<a href="https://agentmods.dev/skills/qoderai/better-harness/generate-harness-dsl"><img src="https://agentmods.dev/badge/skills/qoderai/better-harness/generate-harness-dsl/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/qoderai/better-harness/generate-harness-dsl"><img src="https://agentmods.dev/badge/skills/qoderai/better-harness/generate-harness-dsl.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
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 | $0.00059 | $0.00667 |
| Opus 5.5 | $0.00024 | $0.00267 |
| Sonnet 5.5 | $0.00012 | $0.00133 |
| Haiku 4.5 | $0.00006 | $0.00067 |
Grade A, and why
generate-harness-dsl 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 — 68 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Generate Harness DSL
Create a standalone Harness as Code v0.3 document and prove the execution contract with the package compiler and resolver.
Workflow
- Identify the host, the capabilities it must actually expose, and the real
control owner. Use
sessionfor one host session. Usestate-machineorprogramonly when a selected adapter explicitly implements that mode. - Read the DSL contract. Start from the minimal example; open the standard example when callable tools are required.
- Generate one self-contained document unless the user requests a fragment.
Include
language 0.3, every referenced declaration, a concrete runtime, and a named deployment. Standard tools may remain implicit. - Run
scripts/validate.mjs. Fix every compiler or resolution error and rerun. A compile-only state machine is not an executable Qoder or Pi deployment. - Return the DSL or saved file plus the harness id, deployment id, runtime, resolution status, and any execution boundary the selected adapter cannot satisfy.
Authoring Rules
- State only falsifiable requirements. Do not add permission, setting, degradation, binding, input/output-name, or runtime-execution syntax; v0.3 intentionally has none.
- Match the requirement verb to its capability:
use skill,require tool, orconnect mcp. - Only the standard tool ids in the contract may be undeclared. Every custom
tool declares a stable
contractid that the adapter exposure must match. - Qoder and Pi descriptors run
sessionworkflows only. A session workflow names exactly the one agent role declared by each harness that uses it. - State-machine outcomes are typed on agents. Every route emitter, outcome, destination, entry, and stop must exist, and every agent must be reachable.
- A
program <language> <entry>workflow resolves only when the adapter lists the same language inprogrammaticLanguages. - Keep credentials out of source. Prefer
env.VARIABLEfor MCP endpoints, but remember that declaring an endpoint does not connect it; the adapter must do that. - Do not invoke host SDKs, install integrations, or claim native enforcement while generating or validating DSL.
What ships with it
3 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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 · 68 lines · 59 tokens per session scan A 79b27bc03ad9
generate-harness-dsl is a skill published in the GitHub repository QoderAI/better-harness (2,408 stars, last pushed 4d ago), licensed MIT. It adds 59 tokens to every session and 667 once invoked, about $0.0002 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 2026-08-30.
Other skills, from other repositories
brainstorm
Approach brainstorming: grill a vague requirement into a settled approach, then a user-approved Spec. Takes a depth gear (low = few questions, fast out / medium = the daily default / high = research-grade deep grilling for bottlenecks and innovative design). Triggers: brainstorm, grill, discuss first, spec first…
diagnose
Investigate runtime, build, test, or performance failures whose root cause is unknown. First build a feedback loop that goes red on this failure, then locate the cause via reproduction, falsifiable hypotheses, and targeted instrumentation; a known localized fix needs no full diagnostic process.
plan
For multi-stage tasks, tasks with complex dependencies, or tasks that need cross-session recovery, form an executable plan and, when the project has no recovery surface, establish a minimal recovery surface (workindex + state + AGENTS pointer). When the user only wants a plan, deliver it and stop; when execution is…
review
Review diffs, proposals, or completion claims, checking correctness, risk, and acceptance evidence. Supports limited-scope review of WIP; simple changes can be self-reviewed by implement, and verify is an alias of this skill.
sweep
On-demand repo-wide reconciliation (not a task close-out): recovery-surface detection → inline reconciliation checklist → six-bucket classification → mechanical items auto-fixed, ambiguous items listed with evidence for human adjudication → four retirement steps → report. Triggers: sweep, take stock, big cleanup…
autoresearch
Research-loop controller: the user assigns a goal or open question (the approach may be completely unknown) and this skill runs it independently as a repeated research -> hands-on -> adversarial review -> distill loop — investigate, run experiments, attack your own conclusions, distill; check completion criteria and…