SKILLEMALL.ai
GitHub ↗ RU Sign in

BD harness-engineering

Agent Harness 构建规范——当 AI 通过编程构建 harness agent（coding agent / 自动化 agent / 长时运行 agent）时，读取本 skill 作为参考标准，确保产出符合 harness 工程架构：Loop / Provider / Tools / Permissions / Session / Compaction / Prompts & Skills / Extensions 七层 + Delivery。覆盖设计决策、可运行 Python 代码模板、长时运行模式。触发词：coding agent、agent loop、tool calling、context management、session persistence、agent harness、构建 agent。Use this skill when the user asks to build a coding agent / automation agent / long-running agent / agent harness.

Not recommendedcritical or high security findings
ClawHub Agent Skills author: Edde-101 v0.1.0 MIT-0 7 files body ≈ 854 tokens Open the sourceclawhub.ai↗ analyzed 3 d ago

Agent Harness 构建规范——当 AI 通过编程构建 harness agent（coding agent / 自动化 agent / 长时运行 agent）时，读取本 skill 作为参考标准，确保产出符合 harness 工程架构：Loop / Provider / Tools /…

As a process D 45/100 · Unfinished process — weak spots: result and completion, inputs and preconditions, failures and branches

GeneratorAI and agentstype and topics are labelled automatically from the skill text
Runs in: OpenClaw Claude / Claude Code ChatGPT / Codex Gemini CLI Cursor GitHub Copilot / VS Code Qwen Code Kimi Code CLI Z.ai (GLM) DeepSeek (Deep Code) Mistral Vibe LangChain (deepagents) Goose OpenCode
JSON
Technical rating
B
81/100
safety, quality, tests
Safety 60%
81
Quality 40%
81
Run on models
none yet
Process rating
D
45/100
Unfinished process
Result and completion w 14
0
Inputs and preconditions w 11
0
Failures and branches w 10
0
the three weakest of ten parameters · all ten

What is at stake

The skill contains fragments that, in the wrong hands, cost money or data. Below: what the installer risks and what the author should do.

Dangerous commands
If you install

The skill contains commands that delete files, rewrite disks or execute code fetched from the network. The agent may run them without asking if it believes the instructions require it.

For the author

Replace destructive commands with safe equivalents that ask for confirmation, scope them to one folder, and stop piping curl into a shell: pin a version and a checksum.

How to improve

  1. Address the high-severity findings: each costs 18 safety points. If one is a false positive, add the rule id to guard.allow in spec.yaml.
For the model run — optional
  • Your own cases (evals/evals.json, 4–6 real requests with expected answers): the full check would then run those instead of a model-drafted suite.
  • A spec.yaml with trigger phrases and assertions — a behaviour contract for CI; `skilltest init` writes a template.

Guard findings · 2

  • high Dangerous commands cmd-destructive-fs references/code-templates.md:764
    Destructive filesystem command (wipes root/home/drive) (detector / deny-list definition)
    dangerous = ["rm -rf /", "rm -rf ~", "sudo rm", ":(){:|:&};:"]
    detector
Medium and low: 1
  • low Dangerous commands cmd-privilege references/code-templates.md:764
    Privilege escalation / world-writable permissions (detector / deny-list definition)
    dangerous = ["rm -rf /", "rm -rf ~", "sudo rm", ":(){:|:&};:"]
    detector

Files scanned: 7. Evidence is masked. Grey chips explain why severity was lowered.

Against the Agent Skills spec

  • note frontmatter-key unknown frontmatter key "description_zh"
  • note frontmatter-key unknown frontmatter key "description_en"
  • note frontmatter-key unknown frontmatter key "agent_created"

Process rating: all ten parameters 45/100

  • 0Result and completion. Does not say what the result is
  • 0Inputs and preconditions. Does not say what the process needs to start
  • 0Failures and branches. Linear process with no failure handling
  • 0Progress reporting. Says nothing while it works
  • 30Running it twice. 1 mutating operations with no state check
  • 40Consistency. Frontmatter name (harness-engineering) differs from the folder (harness-engineering-skill)
  • 60Tools and files. Uses tools (bash) that frontmatter does not declare
  • 70When it triggers. States when to use, but not when not to
  • 100Steps. 29 steps
  • 100Execution cost. Instruction body is 854 tokens

Everything here is measured from the skill text rather than judged by a model, so the numbers are checkable. A parameter weighs more when it is a more common reason for the process to stall.

Quality signals

  • +5Description has no quoted example phrases that should trigger the skill
  • +4Description does not say when NOT to use the skill (false activations)
  • +3Output format is not stated: the model decides each time
  • +4No input/output examples
  • +1No license
  • +2Single-language instructions
  • +3Description length 477: enough signal without eating the budget
  • +4Structure: 12 headings
  • +3Step-by-step instructions: 29 items
  • +4Reference files are cited in the instructions (5 of 5)

Quality base 70; lint remarks subtract, signals add up to 100. Result: 81.

External checks

ClawHub: suspicious
The skill is a documentation/template package, but its runnable harness template gives an AI agent broad shell access without enough built-in containment.
LLM: suspicious (high) · 13 Aug 2026