CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

agent-orchestration

AI DevKit · Supervise multi-agent workflows over repeated passes: poll progress, unblock waiting agents, coordinate dependencies, relay outputs, resolve conflicts, and verify completion. Use only for ongoing multi-agent coordination, not one-off list/detail/send/start/kill actions.

Quality

94%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide
SecuritybySnyk

Low

Low-risk findings worth noting

SKILL.md
Quality
Evals
Security

Quality

Content

92%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 exemplary instruction-only skill body: lean, tool-specific where it matters, with a well-sequenced supervision loop and explicit verification gates. The only gap is slightly incomplete command syntax for the detail step and no worked example of a relayed instruction or status report.

Suggestions

Write the detail step as a full command (e.g. "ai-devkit agent detail --tail 10 --id <id>") so it is copy-paste ready.

Include a one-line example of an acceptable brief status report to anchor step 4 of the loop.

Show a minimal template for a self-contained instruction sent to a waiting agent.

DimensionReasoningScore

Conciseness

Every line is operational; no concepts Claude already knows are explained, and non-obvious tool knowledge ("reason: \"timeout\" is normal", "Never use a long sleep: harnesses block or kill it") earns its tokens. Matches the score-5 anchor 'every token earns its place'.

5 / 5

Actionability

Concrete commands are given ("agent list --json", "ai-devkit agent wait --json" with --id/--group guidance) and the action vocabulary is specific (approve, clarify, correct, delegate, relay, verify, escalate). It sits below score 5 because "detail --tail 10" is not a full copy-paste invocation and there is no example of a self-contained instruction or a report status line.

4 / 5

Workflow Clarity

A clear 5-step loop with explicit validation checkpoints ("never assume names/statuses", "Inspect waiting, idle, unknown, missing, or stale agents before acting", "Use verify before accepting any agent's completion claim") and an error-recovery loop for wait timeouts. Matches the score-5 anchor's explicit validation and feedback loops.

5 / 5

Progressive Disclosure

The skill is under 50 lines, needs no external references (no references/, scripts/, or assets/ bundles exist), and is cleanly organized into Rules, Loop, and Completion sections. Per the rubric's simple-skill guidance, this earns a 5.

5 / 5

Total

19

/

20

Passed

Description

92%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 description: concrete third-person action list, explicit two-sided use-conditions with anti-triggers that sharply separate it from sibling skills. The only weakness is modest synonym coverage for natural trigger phrasing.

Suggestions

Add one or two natural synonyms such as "orchestrate" or "monitor agent status" to broaden trigger-term coverage.

Optionally name the 'AI DevKit' context earlier in the trigger clause so users referencing the devkit by another term still match.

DimensionReasoningScore

Specificity

The description lists six concrete actions ("poll progress, unblock waiting agents, coordinate dependencies, relay outputs, resolve conflicts, and verify completion") in third person, comprehensively covering the multi-agent supervision domain. It exceeds the score-4 anchor ('several specific actions, minor gaps') because coverage of the orchestration loop is complete rather than partial.

5 / 5

Completeness

It explicitly answers both what ("Supervise multi-agent workflows over repeated passes" plus the enumerated actions) and when ("Use only for ongoing multi-agent coordination, not one-off list/detail/send/start/kill actions") with concrete trigger and anti-trigger phrases. The explicit 'Use only for...' clause satisfies the score-5 anchor, not merely the score-4 anchor's softer 'could be more specific'.

5 / 5

Trigger Term Quality

Strong natural phrases are present ("multi-agent workflows", "ongoing multi-agent coordination", "unblock waiting agents", "relay outputs"), matching what a user needing this skill would say. It falls short of the score-5 anchor because common synonyms such as "orchestrate agents", "monitor agents", or "agent status" are missing.

4 / 5

Distinctiveness Conflict Risk

The negative exclusion clause ("not one-off list/detail/send/start/kill actions") actively carves out a clear niche from adjacent agent-management and agent-communication skills, giving minimal conflict risk. This is stronger than the score-4 anchor, which still admits minor overlap with closely related skills.

5 / 5

Total

19

/

20

Passed

Validation

100%

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

Validation — 16 / 16 Passed

Validation for skill structure

No warnings or errors.

Repository

codeaholicguy/ai-devkit

Path

skills/agent-orchestration/SKILL.md

Reviewed

Table of Contents

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.