minion-orchestrator

Unified Minions skill for both deterministic shell jobs and LLM subagent orchestration. Replaces the older `gbrain-jobs` routing intent. Use when: submitting gbrain jobs, shell/background tasks, spawning subagents, checking progress, steering running work, pausing/resuming, parallel fan-out. One durable, observable, steerable queue interface. Also carries the durable-execution doctrine for any operation expected to exceed ~2 minutes: capability ladder, deadman checks that verify the result was reported, and content-addressed stage checkpoints for expensive pipelines.

Quality

78%

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/minion-orchestrator/SKILL.md

Quality

Content

77%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.

A highly actionable, well-sequenced skill body with executable commands, explicit validation, and clear feedback loops. The two weaknesses are repetition of the durable-execution doctrine across three sections (conciseness) and the absence of any bundle files with reference-grade appendices inlined rather than split out (progressive disclosure).

Suggestions

Collapse the durable-execution doctrine into one authoritative section and have the Contract and Anti-Patterns reference it instead of restating it, cutting roughly a third of the repeated prose.

Extract the content-addressed-checkpoint appendix (and optionally the flag/timeout reference tables) into a references/ file and link to it from the body to improve progressive disclosure.

Move the admission-control/coalesced detail, which is restated, into the Phase 1 submit section alone and delete the duplicate mention.

Dimension	Reasoning	Score
Conciseness	The body is dense and assumes Claude's intelligence (no "what is a job queue" padding), but the durable-execution doctrine is stated three times — Contract, the full "Durable execution" section, and Anti-Patterns — and the deadman failure modes and coalesced/admission-control notes recur, which is more than the "minor" over-explanation a 4 requires.	3 / 5
Actionability	Provides copy-paste-ready CLI commands ("gbrain jobs submit shell --params '{...}'"), exact MCP calls with parameters ("send_job_message id=ID payload={...}"), source file paths, flag lists, a timeout table, and output-format templates covering the common cases end to end.	5 / 5
Workflow Clarity	Sequences work as Phase 1-5 (Submit → Monitor → Steer → Lifecycle → Review) and the durable-execution ladder as Rung 1→2→3 with explicit preconditions, validation ("Verify setup: run `gbrain jobs stats`"), feedback loops (reported-check → recover/retry), an Eval Contract with hard fails, and checklists ("Sequencing and locks", the timeout table).	5 / 5
Progressive Disclosure	No bundle files exist (references/, scripts/, assets/ all absent), and substantial reference-grade material is inlined in a single ~470-line document — the content-addressed-checkpoint appendix, full flag enumerations, and the timeout table would each justify a separate file; inline source/doc paths are clearly signaled one level deep, but the inlining keeps this below a 4.	3 / 5
	Total	16 / 20 Passed

Description

78%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, specific description with an explicit "Use when:" trigger clause and a clear, low-conflict niche. Its main weakness is entangling the capability/trigger list with durable-execution doctrine prose, which slightly muddies both the "what" and the crispness of the trigger enumeration.

Suggestions

Split the doctrine sentence ("Also carries the durable-execution doctrine...") out of the description or move it after the triggers so the what+when pairing reads cleanly.

Tighten trigger synonyms — add a couple of plain-language variants (e.g., "queue a task", "background command") alongside the technical ones.

Lead with the concrete capability list before the routing/doctrine context so specificity reads as comprehensive rather than interleaved.

Dimension	Reasoning	Score
Specificity	Lists multiple concrete actions across two lanes ("submitting gbrain jobs, shell/background tasks, spawning subagents, checking progress, steering running work, pausing/resuming, parallel fan-out") plus sub-capabilities ("capability ladder, deadman checks... content-addressed stage checkpoints"), but blends them with routing/doctrine prose rather than a clean capability enumeration, stopping short of comprehensive.	4 / 5
Completeness	Explicitly answers both what ("Unified Minions skill for both deterministic shell jobs and LLM subagent orchestration... durable, observable, steerable queue interface") and when (a literal "Use when:" clause with concrete trigger phrases), but the "what" is diluted by the appended durable-execution doctrine, falling just short of the crisp what+when pairing at 5.	4 / 5
Trigger Term Quality	The explicit "Use when:" clause supplies natural phrases users would say ("checking progress", "steering running work", "pausing/resuming", "parallel fan-out"), with good coverage and some synonyms, though a few terms lean technical ("capability ladder", "content-addressed stage checkpoints") and synonym coverage is uneven.	4 / 5
Distinctiveness Conflict Risk	Occupies a clear, narrow niche (Postgres-native durable job queue with two specific lanes) and uses domain-specific triggers ("gbrain jobs", "minions", "deadman") unlikely to fire for another skill; it even distinguishes itself from the sibling it replaces ("Replaces the older `gbrain-jobs` routing intent").	5 / 5
	Total	17 / 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 (526 lines); consider splitting into references/ and linking	Warning
frontmatter_unknown_keys	Unknown frontmatter key(s) found; consider removing or moving to metadata	Warning

	Total	14 / 16 Passed

Repository: garrytan/gbrain
Path: skills/minion-orchestrator/SKILL.md
Commit: 055ac6c

Reviewed: 40 minutes 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.