agent-launcher-orchestrator

Use when a user wants to build, launch, grade, or schedule a Claude Managed Agent (CMA) in their own Anthropic account — "build me an agent", "launch this as a managed agent", "run this on a schedule", "grade my agent against a rubric", "set up a nightly worker". Reads the per-session goal (./my-agent/goal.json), routes deterministically to one of five phase sub-skills (interview → stage-launch → grade-iterate → run-without-you → wrap-up) via goal_router.py, and compiles the goal+phase into an execution shape (single-pass workflow / bounded grade→iterate loop / recurring cron deployment loop) via loop_compiler.py. Forks context so heavy intake (build sheets, payloads, eval cases) stays out of the parent thread. All launches are emitted as BYOK curl the user runs with their own key; no tool makes API calls. Inspired by anthropics/launch-your-agent (Apache-2.0). Distinct from engineering/agent-harness (generic domain loop) and engineering/write-a-skill (authors Claude Code skills, not CMAs).

Quality

94%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

—

The risk profile of this skill

Quality

Content

88%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 actionable, well-sequenced orchestrator body with strong validation checkpoints and feedback loops. The main defect is broken reference links: the body points to reference markdown files that are not present in the bundle.

Suggestions

Create the missing reference files (references/cma-primitives.md, references/loops-and-workflows.md, and interview-to-config.md) or remove/fix the links so navigation resolves.

Drop the duplicate "Inspired by Anthropic's launch-your-agent (Apache-2.0)" line and the restating intro from the body since the description already covers provenance and the through-line.

Consider adding a brief one-line validation note in the hand-off contract that artifacts produced by the sub-skill are checked before `goal_state.py advance` moves the phase.

Dimension	Reasoning	Score
Conciseness	The body is lean and assumes Claude's competence — routing table, pre-flight gates, and forcing-question library all earn their place — but the "Inspired by Anthropic's launch-your-agent" provenance line and the through-line intro restate material already in the description, a minor redundancy that could be trimmed.	4 / 5
Actionability	Provides copy-paste-ready, executable commands with real flags and exit-code semantics — `python3 scripts/goal_router.py --out-dir ./my-agent` with exit 0/3/4 handling, and `loop_compiler.py ... --max-iterations 5 --cron "0 9 * * *" --timezone Europe/Berlin --nest-outcome` — covering the two main operations.	5 / 5
Workflow Clarity	Clear sequence (router → act on exit code → compile loop → fork to sub-skill → advance → digest) with explicit validation checkpoints (pre-flight hard refusals, max_iterations cap 1..20) and feedback loops (exit 3 ASK → forcing question → re-route; exit 4 REFUSE → one sentence → re-route; grade→iterate).	5 / 5
Progressive Disclosure	Well-structured lean overview with clearly signaled one-level-deep references and scripts separated into ./scripts/, but the referenced files do not actually exist — ../../references/cma-primitives.md, ../../references/loops-and-workflows.md, and the cited interview-to-config.md all fail to resolve, breaking navigation.	4 / 5
	Total	18 / 20 Passed

Description

100%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 strong, well-calibrated description that answers what and when with concrete trigger phrases and explicit disambiguation from neighboring skills. Voice is consistently third person, so no specificity penalty applies.

Dimension	Reasoning	Score
Specificity	Lists multiple concrete actions — "build, launch, grade, or schedule", "Reads the per-session goal", "routes deterministically", "compiles the goal+phase into an execution shape", "Forks context", "emitted as BYOK curl" — giving comprehensive coverage of what the skill does.	5 / 5
Completeness	Explicitly answers both halves: a clear "Use when a user wants to build, launch, grade, or schedule a Claude Managed Agent (CMA)..." trigger clause with concrete trigger phrases, plus a detailed account of what the skill does (route, compile, fork, emit BYOK).	5 / 5
Trigger Term Quality	Includes natural phrasings users would actually say — "build me an agent", "launch this as a managed agent", "run this on a schedule", "grade my agent against a rubric", "set up a nightly worker" — alongside the keyword verbs build/launch/grade/schedule.	5 / 5
Distinctiveness Conflict Risk	Occupies a clear niche (CMAs in the user's own Anthropic account) and explicitly disambiguates from siblings: "Distinct from engineering/agent-harness (generic domain loop) and engineering/write-a-skill (authors Claude Code skills, not CMAs)."	5 / 5
	Total	20 / 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
frontmatter_unknown_keys	Unknown frontmatter key(s) found; consider removing or moving to metadata	Warning
relative_links	Relative link issues: 2 suspicious	Warning

	Total	14 / 16 Passed

Repository: alirezarezvani/claude-skills
Path: agent-launcher/skills/agent-launcher-orchestrator/SKILL.md
Commit: 19392f7

Reviewed: ~14 hours 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.