Agent skill

Harness Security Bench

Run @metaharness/darwin security bench (upstream "Darwin Shield" / ADR-155) — evolves a champion security-detection harness against a 10-vuln / 9-decoy corpus and grades it on…

MITAuto-check: notesDevelopment

Install Harness Security Bench

skills CLI

$ npx skills add ruvnet/ruflo --skill harness-security-bench -a claude-code

Project install by default; add -g for ~/.claude/skills/.

GitHub CLI

$ gh skill install ruvnet/ruflo harness-security-bench --agent claude-code

Project scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).

Manual copy

$ git clone --depth 1 https://github.com/ruvnet/ruflo.git skills-src && mkdir -p .claude/skills && cp -r skills-src/plugins/ruflo-metaharness/skills/harness-security-bench .claude/skills/harness-security-bench && rm -rf skills-src

Use ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.

Claude Code skills documentation · loads skills from .claude/skills/

Paste into Claude Code

Install the "harness-security-bench" agent skill from https://github.com/ruvnet/ruflo/tree/main/plugins/ruflo-metaharness/skills/harness-security-bench into .claude/skills/harness-security-bench/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "harness-security-bench", then confirm the skill loads.

Claude Code copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.

$skill-installer

$skill-installer install https://github.com/ruvnet/ruflo/tree/main/plugins/ruflo-metaharness/skills/harness-security-bench

Type this inside Codex. $skill-installer <name> installs a curated skill from openai/skills. The installer writes to $CODEX_HOME/skills (default ~/.codex/skills). Restart Codex if the skill does not show up.

skills CLI

$ npx skills add ruvnet/ruflo --skill harness-security-bench -a codex

Project install goes to .agents/skills/; add -g for ~/.codex/skills/.

GitHub CLI

$ gh skill install ruvnet/ruflo harness-security-bench --agent codex

Project scope by default (.agents/skills/); add --scope user for a personal install.

Manual copy

$ git clone --depth 1 https://github.com/ruvnet/ruflo.git skills-src && mkdir -p .agents/skills && cp -r skills-src/plugins/ruflo-metaharness/skills/harness-security-bench .agents/skills/harness-security-bench && rm -rf skills-src

Use ~/.agents/skills/ instead of .agents/skills for a personal install.

Codex skills documentation · loads skills from .agents/skills/

Paste into Codex

Install the "harness-security-bench" agent skill from https://github.com/ruvnet/ruflo/tree/main/plugins/ruflo-metaharness/skills/harness-security-bench into .agents/skills/harness-security-bench/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "harness-security-bench", then confirm the skill loads.

Codex copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.

skills CLI

$ npx skills add ruvnet/ruflo --skill harness-security-bench -a cursor

Project install goes to .agents/skills/; add -g for ~/.cursor/skills/.

GitHub CLI

$ gh skill install ruvnet/ruflo harness-security-bench --agent cursor

Project scope by default (.agents/skills/); add --scope user for a personal install.

Manual copy

$ git clone --depth 1 https://github.com/ruvnet/ruflo.git skills-src && mkdir -p .cursor/skills && cp -r skills-src/plugins/ruflo-metaharness/skills/harness-security-bench .cursor/skills/harness-security-bench && rm -rf skills-src

Use ~/.cursor/skills/ instead of .cursor/skills for a personal install.

Cursor skills documentation · loads skills from .cursor/skills/, .agents/skills/, .claude/skills/, .codex/skills/

Paste into Cursor

Install the "harness-security-bench" agent skill from https://github.com/ruvnet/ruflo/tree/main/plugins/ruflo-metaharness/skills/harness-security-bench into .cursor/skills/harness-security-bench/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "harness-security-bench", then confirm the skill loads.

Cursor copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.

gemini skills install

$ gemini skills install https://github.com/ruvnet/ruflo.git --path plugins/ruflo-metaharness/skills/harness-security-bench

--scope user (default) or --scope workspace; --path is the subfolder of the repo that holds the skill; --consent skips the security confirmation prompt.

skills CLI

$ npx skills add ruvnet/ruflo --skill harness-security-bench -a gemini-cli

Project install goes to .agents/skills/; add -g for ~/.gemini/skills/.

GitHub CLI

$ gh skill install ruvnet/ruflo harness-security-bench --agent gemini-cli

Project scope by default (.agents/skills/); add --scope user for a personal install.

Manual copy

$ git clone --depth 1 https://github.com/ruvnet/ruflo.git skills-src && mkdir -p .gemini/skills && cp -r skills-src/plugins/ruflo-metaharness/skills/harness-security-bench .gemini/skills/harness-security-bench && rm -rf skills-src

Use ~/.gemini/skills/ instead of .gemini/skills for a personal install, then run /skills reload.

Gemini CLI skills documentation · loads skills from .gemini/skills/, .agents/skills/

Paste into Gemini CLI

Install the "harness-security-bench" agent skill from https://github.com/ruvnet/ruflo/tree/main/plugins/ruflo-metaharness/skills/harness-security-bench into .gemini/skills/harness-security-bench/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "harness-security-bench", then confirm the skill loads.

Gemini CLI copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.

GitHub CLI

$ gh skill install ruvnet/ruflo harness-security-bench

Installs for Copilot at project scope by default; add --scope user for a personal install. Preview a skill first with gh skill preview. Needs GitHub CLI 2.90.0 or later (public preview).

skills CLI

$ npx skills add ruvnet/ruflo --skill harness-security-bench -a github-copilot

Project install goes to .agents/skills/; add -g for ~/.copilot/skills/.

Manual copy

$ git clone --depth 1 https://github.com/ruvnet/ruflo.git skills-src && mkdir -p .github/skills && cp -r skills-src/plugins/ruflo-metaharness/skills/harness-security-bench .github/skills/harness-security-bench && rm -rf skills-src

Use ~/.copilot/skills/ instead of .github/skills for a personal install. Commit .github/skills so cloud agent and code review can use it.

GitHub Copilot skills documentation · loads skills from .github/skills/, .claude/skills/, .agents/skills/

Paste into GitHub Copilot

Install the "harness-security-bench" agent skill from https://github.com/ruvnet/ruflo/tree/main/plugins/ruflo-metaharness/skills/harness-security-bench into .github/skills/harness-security-bench/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "harness-security-bench", then confirm the skill loads.

GitHub Copilot copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.

skills CLI

$ npx skills add ruvnet/ruflo --skill harness-security-bench -a opencode

OpenCode documents no install command of its own. Project install goes to .agents/skills/; add -g for ~/.config/opencode/skills/.

GitHub CLI

$ gh skill install ruvnet/ruflo harness-security-bench --agent opencode

Project scope by default (.agents/skills/); add --scope user for a personal install.

Manual copy

$ git clone --depth 1 https://github.com/ruvnet/ruflo.git skills-src && mkdir -p .opencode/skills && cp -r skills-src/plugins/ruflo-metaharness/skills/harness-security-bench .opencode/skills/harness-security-bench && rm -rf skills-src

Use ~/.config/opencode/skills/ instead of .opencode/skills for a personal install.

OpenCode skills documentation · loads skills from .opencode/skills/, .claude/skills/, .agents/skills/

Paste into OpenCode

Install the "harness-security-bench" agent skill from https://github.com/ruvnet/ruflo/tree/main/plugins/ruflo-metaharness/skills/harness-security-bench into .opencode/skills/harness-security-bench/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "harness-security-bench", then confirm the skill loads.

OpenCode copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.

Facts

Skill name: harness-security-bench
GitHub stars: 74k
Token cost: ~1.1k tokens
SKILL.md length: 357 words
Files: 1
Skills in repo: 264
Repo updated: Oct 9, 2026
First seen: Oct 7, 2026
Licence: MIT

At a glance

Run @metaharness/darwin security bench (upstream "Darwin Shield" / ADR-155) — evolves a champion security-detection harness against a 10-vuln / 9-decoy corpus and grades it on…

Works in 5 steps: Run metaharness-darwin security bench… → Default timeout = 3s × 19 evaluations ×… → Parse the markdown report — overall… → …
Tasks that involve Architecture decision records
SKILL.md covers Why this matters for ruflo's…, Algorithm, Output shape and Wiring into ADR-155 nightly…, plus 2 more sections
Instructions only: no scripts, shell commands, URLs or credentials in SKILL.md

What it does

Harness Security Bench is an agent skill from ruvnet/ruflo. Run @metaharness/darwin security bench (upstream "Darwin Shield" / ADR-155) — evolves a champion security-detection harness against a 10-vuln / 9-decoy corpus and grades it on TPR/FPR/patch-pass/repro/unsafe vs four baselines (B0 static, B1 LLM-single-pass, B2 fixed-agent, B3 Darwin-champion). Closest reference implementation for ruflo's own ADR-155 nightly self-learning security harness (PR

Its SKILL.md is about 1.1k tokens, which your agent loads only when the skill is triggered. It is a single SKILL.md file with no bundled scripts.

It sits in Development, covering Architecture decision records. The repository describes itself as: 🌊 The original agent harness. Deploy intelligent multi-player swarms, coordinate autonomous workflows, and build conversational AI systems. Features adaptive memory…. The licence is MIT.

When your agent uses it

Tasks that involve Architecture decision records

Example prompts

“Darwin Shield”
“/harness-security-bench”

Requirements

Pre-approved tools (allowed-tools): Bash

Workflow steps

5 steps, taken from the first numbered list in SKILL.md.

Run metaharness-darwin security bench --population N --cycles N [--seed S] from the
Default timeout = 3s × 19 evaluations × population × cycles + 30s overhead.
Parse the markdown report — overall PASS/FAIL plus per-gate
Parse the baselines-vs-champion table (4 rows: fitness/TPR/FPR/patchPass/
Emit structured JSON. With --alert-on-fail, exit 1 when overall = FAIL.

What it can do on your machine

Read from SKILL.md and the folder at commit 58e0ae7. It shows what the files ask for, not the result of running them.

Tool permissions
Pre-approves these tools, so the agent can use them without asking each time:
- Bash
From allowed-tools in the SKILL.md frontmatter.
Runs code
No scripts in the folder and no shell commands in SKILL.md (its code samples are json and jsonc).
From the folder's file list and the shell code blocks in SKILL.md.
Network
No URLs in SKILL.md.
From URLs in SKILL.md, links to its own repository left out.
Credentials
Names no API keys, tokens, secrets or passwords.
From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Harness Security Bench loads about 1.1k tokens when it runs. Until then it costs about 105 tokens; SKILL.md has 357 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it: ~105
When it runs · the whole SKILL.md, loaded when a task matches: ~1.1k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Auto-check: notes

The automated check noted patterns worth knowing about, such as sudo or a known installer.

NotePre-approves every shell command (allowed-tools: Bash)SKILL.md
```
allowed-tools: Bash
```

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.

SKILL.md

The full file from ruvnet/ruflo at commit 58e0ae7, republished under its MIT licence (© ruvnet). 357 words, ~1,133 tokens.

Download SKILL.mdSave it as .claude/skills/harness-security-bench/SKILL.md (or your agent's skills folder).

name: harness-security-bench
description: Run `@metaharness/darwin security bench` (upstream "Darwin Shield" / ADR-155) — evolves a champion security-detection harness against a 10-vuln / 9-decoy corpus and grades it on TPR/FPR/patch-pass/repro/unsafe vs four baselines (B0 static, B1 LLM-single-pass, B2 fixed-agent, B3 Darwin-champion). Closest reference implementation for ruflo's own ADR-155 nightly self-learning security harness (PR
allowed-tools: Bash
argument-hint: [--population 2] [--cycles 1] [--seed N] [--alert-on-fail]

Surfaces the upstream metaharness-darwin security bench command. This is the upstream's own ADR-155 — Darwin Shield — and is the closest reference implementation for ruflo's nightly self-learning security harness (#2417).

Why this matters for ruflo's ADR-155

ruflo's ADR-155 proposes three learning loops (per-dimension confidence, severity calibration, auto-fix bid). Loop A trains on accumulated (finding, dimension, human_outcome) tuples — but the gradient signal is only sound if the underlying detection mechanism converges on a known-good corpus. Darwin Shield evolves exactly that mechanism on a 10-vuln/9-decoy ground-truth set. Running this nightly gives us:

Empirical floor: if Darwin Shield's champion can't reach TPR=1/FPR=0 on the bench corpus, our Loop A's reward signal is noise.
Drift detection: week-over-week champion fitness deltas surface when the security landscape (or our mutator policy) shifts.
Baseline diversity: the 4 baselines (B0–B3) give us 4 anchor points to weight per-dimension confidence against.

Algorithm

Implementation: scripts/security-bench.mjs.

Run metaharness-darwin security bench --population N --cycles N [--seed S] from the installed @metaharness/darwin (pin in _darwin.mjs; one-time cache install only if no installed copy qualifies — never npx).
Default timeout = 3s × 19 evaluations × population × cycles + 30s overhead. At default --population 2 --cycles 1 ≈ 144s; at --population 4 --cycles 3 ≈ 12 min.
Parse the markdown report — overall PASS/FAIL plus per-gate pass/fail rows (gate examples: "TPR improvement ≥ 25% vs fixed", "FPR reduction ≥ 40%", "Patch-test pass rate ≥ 80%", "Reproduction success ≥ 90%", "Unsafe outputs = 0", "Cost increase ≤ 2× fixed", "Beyond SOTA: champion statistically beats previous champion", "Compounding: false-positive repeat-rate drop ≥ 35%").
Parse the baselines-vs-champion table (4 rows: fitness/TPR/FPR/patchPass/ repro/unsafe/cost per harness).
Emit structured JSON. With --alert-on-fail, exit 1 when overall = FAIL.

Show full SKILL.md (92 more words)Show less

Output shape

json

{
  "success": true,
  "data": {
    "overall": { "ok": true, "icon": "✅" },
    "gates": {
      "total": 11,
      "passed": 11,
      "failed": 0,
      "details": [{ "ok": true, "criterion": "TPR improvement ≥ 25% vs fixed harness", "measured": "+150% (B2 0.4 → B3 1)" }, ...]
    },
    "baselines": [
      { "harness": "static-only", "fitness": 0.5665, "tpr": 0.3, "fpr": 1, "unsafe": 0, ... },
      { "harness": "LLM single-pass", "fitness": 0.1365, ... },
      { "harness": "fixed agent", "fitness": 0.598, ... },
      { "harness": "Darwin champion", "fitness": 0.93275, "tpr": 1, "fpr": 0, ... }
    ],
    "rawMarkdown": "...",
    "shape": { "population": 2, "cycles": 1, "seed": null },
    "durationMs": 142000
  }
}

Wiring into ADR-155 nightly harness

The ADR-155 nightly workflow (per #2418 task W1.5) will spawn this as one of the active-pentest dimension's calls — its results become a trajectory record:

jsonc

{
  "dimension": "mcp-pentest",
  "subdimension": "darwin-shield-bench",
  "champion_fitness": 0.93275,
  "champion_tpr": 1, "champion_fpr": 0,
  "gates_passed": 11, "gates_failed": 0,
  "shape": { "population": 4, "cycles": 3 }
}

Loop A learns: if darwin-shield-bench consistently passes on the seeded corpus, weight findings caught only by mcp-pentest higher.

Exit codes

Code	Meaning
0	Bench ran (overall PASS or FAIL — distinguish via JSON `overall.ok`), or degraded
1	`--alert-on-fail` and `overall.ok === false`
2	Config error or upstream infrastructure failure

Graceful degradation

When @metaharness/darwin is absent, emits {degraded: true, reason: 'metaharness-darwin-not-available'} and exits 0.

Files

Just SKILL.md in plugins/ruflo-metaharness/skills/harness-security-bench of ruvnet/ruflo.

Open the folder on GitHubat commit 58e0ae7

Compare with similar skills

Harness Security Bench next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.

Harness Security Bench compared with similar skills
Skill	Stars	Used in	Tokens	Auto-check	Licence	Repo updated
Harness Security Bench this skillruvnet/ruflo	74k	—	~1.1k	Automated safety check: Notes	MIT	today
PR Design DocOpenHands/OpenHands	90k	—	~2.4k	Automated safety check: Pass	MIT	today
Cto AdvisorIbrahim-3d/orchestrator-supaconductor	381	4 repos	~2.4k	Automated safety check: Pass	MIT	11 days ago
Architecture DecisionDonchitos/Claude-Code-Game-Studios	26k	—	~1.7k	Automated safety check: Pass	MIT	yesterday
Improve Codebase Architectureywwynm/EverythingDone	144	15 repos	~1.3k	Automated safety check: Pass	GPL-3.0	today
Domain Modelingbrim-borium/spotify_sdk	166	5 repos	~806	Automated safety check: Pass	Apache-2.0	today

Similar skills

PR Design Doc
OpenHands/OpenHands
For a non-trivial pull request, write a self-contained HTML design doc under the temporary .pr/ directory and link a visibility-appropriate preview in the PR description, so maintainers grasp the…
90k GitHub stars~2.4k tokensUpdated today
DevelopmentAuto-check passed
Cto Advisor
Ibrahim-3d/orchestrator-supaconductor
Technical leadership guidance for engineering teams, architecture decisions, and technology strategy.
381 GitHub starsUsed in 4 repos~2.4k tokens
DevelopmentAuto-check passed
Architecture Decision
Donchitos/Claude-Code-Game-Studios
Create an ADR documenting a technical decision: context, alternatives considered, consequences.
26k GitHub stars~1.7k tokensUpdated yesterday
DevelopmentAuto-check passed
Improve Codebase Architecture
ywwynm/EverythingDone
Find deepening opportunities in a codebase, informed by the domain language in CONTEXT.md and the decisions in docs/adr/.
144 GitHub starsUsed in 15 repos~1.3k tokens
DevelopmentAuto-check passed
Domain Modeling
brim-borium/spotify_sdk
Build and sharpen a project's domain model. An agent skill from brim-borium/spotify_sdk.
166 GitHub starsUsed in 5 repos~806 tokens
DevelopmentAuto-check passed
Design Doc Mermaid
SpillwaveSolutions/design-doc-mermaid
Create Mermaid diagrams (flowchart, sequence, class, ER, state, C4, architecture) from text or source code.
176 GitHub starsUsed in 1 repo~5.6k tokens
DevelopmentAuto-check passed

Questions about Harness Security Bench

What does Harness Security Bench do?

Run @metaharness/darwin security bench (upstream "Darwin Shield" / ADR-155) — evolves a champion security-detection harness against a 10-vuln / 9-decoy corpus and grades it on…. Harness Security Bench is an agent skill from ruvnet/ruflo. Run @metaharness/darwin security bench (upstream "Darwin Shield" / ADR-155) — evolves a champion security-detection harness against a 10-vuln / 9-decoy corpus and grades it on TPR/FPR/patch-pass/repro/unsafe vs four baselines (B0 static, B1 LLM-single-pass, B2 fixed-agent, B3 Darwin-champion).

When should I use Harness Security Bench?

Harness Security Bench fits situations like: tasks that involve Architecture decision records.

How do I install Harness Security Bench in Claude Code?

Run `npx skills add ruvnet/ruflo --skill harness-security-bench -a claude-code`. Or copy the skill folder (plugins/ruflo-metaharness/skills/harness-security-bench in ruvnet/ruflo) into .claude/skills/harness-security-bench in your project. Claude Code loads it when a task matches its description.

How do I install Harness Security Bench in Codex?

Run `npx skills add ruvnet/ruflo --skill harness-security-bench -a codex`. Or copy the skill folder (plugins/ruflo-metaharness/skills/harness-security-bench in ruvnet/ruflo) into .agents/skills/harness-security-bench in your project. Codex loads it when a task matches its description.

Can I use Harness Security Bench in Cursor, Gemini CLI or GitHub Copilot?

Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add ruvnet/ruflo --skill harness-security-bench -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/harness-security-bench, .gemini/skills/harness-security-bench, .github/skills/harness-security-bench and .opencode/skills/harness-security-bench in your project.

What does Harness Security Bench need to run?

SKILL.md names no scripts, command-line tools or credentials: Harness Security Bench is instructions for the agent only. Its frontmatter pre-approves these tools: Bash.

Does Harness Security Bench access the network?

SKILL.md contains no URLs. Any network use would come from the scripts or tools the agent runs. This is read from the text; nothing was executed.

Is Harness Security Bench safe to install?

Our automated static check of SKILL.md found notes only (pre-approves every shell command (allowed-tools: bash)), nothing it rates as a warning. It is not a guarantee. Review the folder before installing.

What licence does Harness Security Bench use?

Harness Security Bench is published under the MIT licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does Harness Security Bench use?

About 1.1k tokens (SKILL.md is roughly 4.5k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full.

What are the alternatives to Harness Security Bench?

Skills that share tags, products or a category with Harness Security Bench: PR Design Doc (OpenHands/OpenHands, 90k stars), Cto Advisor (Ibrahim-3d/orchestrator-supaconductor, 381 stars), Architecture Decision (Donchitos/Claude-Code-Game-Studios, 26k stars) and Improve Codebase Architecture (ywwynm/EverythingDone, 144 stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Harness Security Bench?

ruvnet (a GitHub user) maintains it in ruvnet/ruflo, which has 74,159 GitHub stars. The repository holds 264 skills in this directory. The repository was last updated on October 9, 2026.

Source: ruvnet/ruflo on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.

Harness Security Bench

Install Harness Security Bench

Facts

At a glance

What it does

When your agent uses it

Example prompts

Requirements

Workflow steps

What it can do on your machine

Tool permissions

Runs code

Network

Credentials

Context cost

Safety

SKILL.md

Why this matters for ruflo's ADR-155

Algorithm

Output shape

Wiring into ADR-155 nightly harness

Exit codes

Graceful degradation

Files

Compare with similar skills

Similar skills

PR Design Doc

Cto Advisor

Architecture Decision

Improve Codebase Architecture

Domain Modeling

Design Doc Mermaid

More from ruvnet/ruflo

AgentDB Pattern Memory

Security Audit Scanner

SPARC Development Methodology

Swarm Orchestration

Ruflo Multi-Agent Orchestration

Agent Coordination

Questions about Harness Security Bench

Install Harness Security Bench

Facts

At a glance

What it does

When your agent uses it

Example prompts

Requirements

Workflow steps

What it can do on your machine

Tool permissions

Runs code

Network

Credentials

Context cost

Safety

SKILL.md

Why this matters for ruflo's ADR-155

Algorithm

Output shape

Wiring into ADR-155 nightly harness

Exit codes

Graceful degradation

Files

Compare with similar skills

Similar skills

More from ruvnet/ruflo

Related

Questions about Harness Security Bench