launch-openshell-gator

Launch and supervise OpenShell gator agents. Use when starting gator on issues or PRs, checking gator sandboxes, building the gator sandbox image, restarting stuck gators, inspecting gator logs, or experimenting with gator harness/model overrides. Trigger keywords - launch gator, start gator, run gator, gator sandbox, supervised gator, gator logs, restart gator.

Quality

90%

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

Quality

Content

81%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 thorough, highly actionable operator runbook with executable commands, a validated multi-step preflight, and clear feedback loops for destructive restarts. Its main weakness is conciseness: the same input-normalization guards are inlined repeatedly, and the long body is not split into reference files.

Suggestions

Factor the repeated 'gateway_name/pr_number/sandbox_name' validation guards into a single sourced shell helper and call it from each example to remove ~30-40 lines of duplication.

Move the sentinel-interpretation tables and the Troubleshooting section into a reference file (e.g. references/troubleshooting.md) referenced from SKILL.md to reduce the body length and improve progressive disclosure.

Consider a short 'Standard launch' quick-start block at the top that links to the variant-specific sections, so operators can reach the common path without scrolling past the full preflight.

Dimension	Reasoning	Score
Conciseness	Prose is lean and assumes Claude's competence (no lectures on git/gh/Docker basics), but the identical input-normalization guard blocks ('[[ "$gateway_name" =~ ^[A-Za-z0-9_.-]+$ ]]...') are repeated inline across roughly a dozen code blocks and could be factored into a shared helper.	3 / 5
Actionability	Provides fully executable, copy-paste-ready commands covering the common cases — preflight checks, full './scripts/agents/run.sh' invocations with every flag, and concrete prompt strings for PR, issue, paired, authorized, and experimental launches.	5 / 5
Workflow Clarity	The Preflight is a clearly sequenced 6-step procedure with per-step validation and explicit feedback loops (auth fail -> refresh; gateway dead -> debug-openshell-cluster), and the destructive restart path validates staleness before deleting, satisfying the destructive-operation checkpoint.	5 / 5
Progressive Disclosure	Well-organized with clear headers, a Key Paths table, and easy navigation, but the ~410-line body is monolithic with no reference files splitting out the sentinel/tables/troubleshooting material, so it is good rather than optimally split.	4 / 5
	Total	17 / 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 highly specific, third-person description that explicitly states capabilities, gives a concrete 'Use when...' trigger clause, and lists natural trigger keywords with synonyms. It cleanly hits the top anchor on every dimension with no fluff or over-claims.

Dimension	Reasoning	Score
Specificity	Lists multiple concrete operator actions — 'starting gator on issues or PRs, checking gator sandboxes, building the gator sandbox image, restarting stuck gators, inspecting gator logs, or experimenting with gator harness/model overrides' — giving comprehensive coverage rather than a few generic verbs.	5 / 5
Completeness	Clearly answers 'what' ('Launch and supervise OpenShell gator agents') and 'when' via an explicit 'Use when...' clause with concrete trigger phrases, matching the top anchor.	5 / 5
Trigger Term Quality	An explicit 'Trigger keywords -' list supplies natural phrases a user would say (launch gator, start gator, run gator, gator sandbox, supervised gator, gator logs, restart gator), including synonyms (launch/start/run).	5 / 5
Distinctiveness Conflict Risk	The niche is narrow and tool-specific ('OpenShell gator agents', gator sandboxes) with gator-scoped trigger keywords, so it is unlikely to fire for unrelated skills.	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
relative_links	Relative link issues: 9 missing	Warning
referenced_paths_exist	Referenced path issues: 19 missing, 19 deeper-than-1-level	Warning

	Total	14 / 16 Passed

Repository: NVIDIA/OpenShell
Path: .agents/skills/launch-openshell-gator/SKILL.md
Commit: 69a05eb

Reviewed: 1 day 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.