acpx

Use acpx as a headless ACP CLI for agent-to-agent communication, including installed-agent inspection, prompt/exec/sessions workflows, session scoping, queueing, permissions, output formats, system-prompt overrides, and multi-agent flows authored with defineFlow/decision/decisionEdge.

Quality

75%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Low

Low-risk findings worth noting

Fix and improve this skill with Tessl

tessl review fix ./skills/acpx/SKILL.md

Quality

Content

71%Weight 40%Scale 1-5

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

An exceptionally actionable, accurate CLI reference with copy-paste-ready commands and full coverage of session, permission, and flow workflows. Its weaknesses are length and monolithic structure: embedded-host API details, queue internals, and edge-case model handling should be split into one-level-deep reference files, and the description lacks trigger guidance.

Suggestions

Split the 'Embedded hosts' section, queue-ownership internals, and model-catalog/bracketed-ID rules into one-level-deep reference files (e.g. references/embedded.md, references/sessions.md) linked from concise overview sections.

Trim or tabulate the dense edge-case prose (e.g. the 'Cancel / Mode / Config / Model' behavior bullets) to the rules a user needs before running a command, moving exception handling to a reference.

Add explicit verification steps for riskier workflows (e.g. 'sessions show' after import, checking run artifacts under ~/.acpx/flows/runs/<runId>/ after a flow run) to strengthen validation checkpoints.

Dimension	Reasoning	Score
Conciseness	The body is dense and fluff-free with no explanations of concepts Claude already knows, but ~680 lines of inline edge-case minutiae (embedded-host API semantics, Cursor bracketed model ID rules, queue lease-file internals, permission escalation _meta details) could be significantly tightened or moved to reference material. Not a 2 because nearly every sentence carries unique information rather than padding.	3 / 5
Actionability	Fully executable throughout: complete command grammar, copy-paste bash examples for every workflow (install, ensure, prune, system-prompt override, JSON piping), concrete config keys and file paths, and a complete runnable TypeScript flow example. Specific examples cover the common cases.	5 / 5
Workflow Clarity	Commands are clearly sequenced (install -> sessions ensure -> prompt -> output format) and the destructive/batch 'sessions prune' operation has an explicit validation checkpoint ('--dry-run previews what would be deleted without touching disk', shown first in examples), so the destructive-operation cap does not apply. Minor gaps: no explicit verification steps for import, compare, or flow runs.	4 / 5
Progressive Disclosure	Section structure is clear with descriptive headers and outbound links to per-agent guides and docs, but no bundle files exist and the single monolithic file inlines reference-grade content (embedded hosts, queue ownership internals, model-catalog subtleties) that clearly belongs in separate one-level-deep reference files. References are remote URLs rather than well-signaled bundle files.	3 / 5
	Total	15 / 20 Passed

Description

70%Weight 40%Scale 1-5

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

A specific, well-scoped description that clearly communicates what acpx does across its full capability surface. Its main weakness is the absence of any 'Use when...' trigger clause, which limits discoverability and caps completeness.

Suggestions

Add an explicit trigger clause, e.g. 'Use when you need to run coding agents (codex, claude, gemini) headlessly from scripts or the CLI, manage persistent agent sessions, or orchestrate multi-agent workflows via ACP.'

Include natural user phrasings and agent names (e.g. 'run codex/claude from the command line', 'agent sessions', 'headless agent runs') to improve trigger-term coverage.

Consider naming a few concrete actions (queue prompts, cancel in-flight turns, prune stale sessions) instead of umbrella categories like 'permissions' and 'output formats'.

Dimension	Reasoning	Score
Specificity	The description names the tool and lists many concrete capability areas ('installed-agent inspection, prompt/exec/sessions workflows, session scoping, queueing, permissions, output formats, system-prompt overrides'), sitting between anchors 4 and 5. It falls short of anchor 5 because several items are workflow names rather than concrete actions, and capabilities like cancel, auth, and status are folded into umbrella terms.	4 / 5
Completeness	The 'what' is clear and detailed ('headless ACP CLI for agent-to-agent communication, including...'), but there is no 'Use when...' clause or any equivalent explicit trigger guidance, which caps completeness at 3 per the rubric guideline. It is not a 2 because the 'what' half is concrete and comprehensive.	3 / 5
Trigger Term Quality	Good natural keyword coverage including 'acpx', 'ACP', 'agent-to-agent communication', 'sessions', and 'multi-agent flows'. Not anchor 5 because common user phrasings are missing, such as specific agent names (codex, claude) or synonyms like 'headless agent'/'run agents from scripts'.	4 / 5
Distinctiveness Conflict Risk	'acpx', 'ACP', 'agent-to-agent', and 'defineFlow/decision/decisionEdge' anchor a clear niche with distinct triggers and minimal conflict risk; generic words like 'permissions' and 'sessions' are subordinate to the distinct tool identity.	5 / 5
	Total	16 / 20 Passed

Validation

87%

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 — 14 / 16 Passed

Validation for skill structure

Criteria	Description	Result
skill_md_line_count	SKILL.md is long (687 lines); consider splitting into references/ and linking	Warning
relative_links	Relative link issues: 1 suspicious	Warning

	Total	14 / 16 Passed

Repository: openclaw/acpx
Path: skills/acpx/SKILL.md
Commit: 27efb1b

Reviewed: 2 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.