harness-agents

Add or use full agent harness runtimes like Claude Code, Codex, Pi, Cursor, Mastra, or ACP agents inside Agent Native.

Quality

68%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Passed

No findings from the security scan

Fix and improve this skill with Tessl

tessl review fix ./.agents/skills/harness-agents/SKILL.md

Quality

Content

65%

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

The content is highly actionable with executable code and specific identifiers, but it is a long monolithic document with no reference-file split and a setup workflow that lacks validation checkpoints. Splitting sections and adding verification steps would lift the weaker dimensions.

Suggestions

Move the ACP Agents, Code Execution Sandbox, and Sub-Agent Delegation Depth sections into reference files under references/ and link to them from a concise overview, so progressive disclosure reaches one-level-deep navigation.

Add an explicit validation/checkpoint step to the 'How' workflow (e.g. verify the harness session row exists after startAgentHarnessRun, or confirm goalId=agent-harness projection before proceeding).

Deduplicate the install-hint and action-exposure guidance that appears in both 'Adapter Guidance' and 'Don't' to tighten conciseness.

Dimension	Reasoning	Score
Conciseness	The body uses terse directive bullets and explains no general concepts Claude already knows, but at ~160 lines there is thematic overlap between 'Adapter Guidance' and 'Don't' (e.g. install hints, action exposure) that could be tightened.	2 / 3
Actionability	Two complete, executable TypeScript blocks with real imports and signatures, plus specific function names (saveAgentHarnessSession, createAgentHarnessBackgroundAgentController), env vars (AGENT_NATIVE_SANDBOX, AGENT_NATIVE_MAX_SUBAGENT_DEPTH), and file paths make it copy-paste ready.	3 / 3
Workflow Clarity	The 'How' section gives a clear 4-step sequence (register/resolve, start turn, persist state, surface runs) but has no validation checkpoints, matching 'steps listed but validation gaps'.	2 / 3
Progressive Disclosure	Sections are well organized, but the file is a monolithic ~160 lines (above the 50-line simple-skill exception) with ACP, sandbox, and sub-agent content inline that could be split into reference files; no bundle files exist.	2 / 3
	Total	9 / 12 Passed

Description

72%

Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.

The description is specific and distinctive with strong natural trigger terms, but it lacks an explicit 'Use when...' clause, so completeness is capped at 2. Adding trigger guidance would raise the overall score.

Suggestions

Append a 'Use when...' clause naming the situations that should activate this skill (e.g. 'Use when integrating Claude Code, Codex, Cursor, Mastra, Pi, or ACP agents into Agent Native').

Broaden the action verbs beyond 'add or use' to enumerate the concrete operations the skill covers (e.g. register, resolve, run, persist, and surface harness sessions).

Dimension	Reasoning	Score
Specificity	Names concrete targets ('Claude Code, Codex, Pi, Cursor, Mastra, or ACP agents') and two actions ('Add or use'), but the action variety is thin compared to the anchor-3 example of three distinct verbs, so it sits at 'names domain and some actions, not comprehensive'.	2 / 3
Completeness	The 'what' is explicit ('Add or use full agent harness runtimes... inside Agent Native'), but there is no 'Use when...' clause or equivalent trigger guidance, which the guidelines cap at 2.	2 / 3
Trigger Term Quality	Real product names (Claude Code, Codex, Cursor, Mastra) are natural terms a user would actually say when needing this skill; 'ACP agents' and 'Agent Native' are technical but domain-appropriate, giving good coverage.	3 / 3
Distinctiveness Conflict Risk	A narrow, well-defined niche ('agent harness runtimes... inside Agent Native') with distinct named triggers makes it unlikely to fire for the wrong skill.	3 / 3
	Total	10 / 12 Passed

Validation

93%

Warnings & errors only

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation — 15 / 16 Passed

Validation for skill structure

Criteria	Description	Result
frontmatter_unknown_keys	Unknown frontmatter key(s) found; consider removing or moving to metadata	Warning

	Total	15 / 16 Passed

Repository: BuilderIO/agent-native
Path: .agents/skills/harness-agents/SKILL.md
Commit: c1ee18b

Reviewed: 5 days ago

Table of Contents

Discovery Implementation Validation

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.