runpod-compute

Provision and manage GPU pods on RunPod for explicitly chosen long-running research experiments. Use when a Feynman replication, benchmark, or dataset-heavy research run needs persistent GPU compute with SSH access.

Quality

85%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

High

Do not use without reviewing

Quality

Content

87%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 concise, highly actionable command reference with clean section structure and no bundle dependencies. Its main gap is the absence of an explicit lifecycle workflow with validation before destructive operations like pod removal.

Suggestions

Add a short lifecycle workflow (create pod → SSH in → stop/remove) with an explicit validation checkpoint before `runpodctl remove pod`, e.g. confirm the pod id via `runpodctl get pod` and verify no running workloads before terminating.

Before destructive removal, include a verification step such as checking pod status or prompting to confirm the target id, which would lift workflow_clarity above the destructive-operations cap of 3.

Optionally note cost/hygiene validation (e.g. confirm pod is stopped before removal, or that volumes are no longer needed) to make the stop→remove transition safer.

Dimension	Reasoning	Score
Conciseness	The body is a lean ~40 lines of tables, lists, and short command snippets with no concept explanations or padding; the opening sentence orients and sets a useful boundary, and every section earns its place, matching the 'lean and efficient, every token earns its place' anchor.	5 / 5
Actionability	The commands table gives fully specified, copy-paste-ready invocations (e.g. a complete `create pod` with real GPU type and image name) plus executable setup and SSH snippets, with dynamic placeholders explicitly sourced from `runpodctl get pod`, matching the top anchor.	5 / 5
Workflow Clarity	The content is organized as a command reference with only an implicit create→SSH→stop/remove sequence, and the destructive `remove pod` (terminate and delete) operation has no validation checkpoint, so per the destructive-operations cap workflow clarity cannot exceed 3.	3 / 5
Progressive Disclosure	With no bundle files and under 50 lines, the skill relies on well-organized section headers (Setup, Commands, SSH access, GPU types, When to use) for navigation, which satisfies the simple-skill exception for a top progressive-disclosure score.	5 / 5
	Total	18 / 20 Passed

Description

82%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 tight, third-person description that explicitly answers both what and when with concrete, domain-specific trigger phrases and minimal conflict risk. Its only weakness is that the capability list is limited to two verbs rather than a fuller enumeration of operations.

Dimension	Reasoning	Score
Specificity	"Provision and manage GPU pods on RunPod" names the domain with two concrete actions (provision, manage) but does not enumerate several specific operations, matching the '1-2 concrete actions, not comprehensive' anchor and falling short of the 'several specific actions' level above.	3 / 5
Completeness	It clearly states what ("Provision and manage GPU pods on RunPod") and gives an explicit 'Use when a Feynman replication, benchmark, or dataset-heavy research run needs persistent GPU compute with SSH access' clause with concrete triggers, matching the top anchor.	5 / 5
Trigger Term Quality	"Feynman replication, benchmark, or dataset-heavy research run", "persistent GPU compute with SSH access" give good natural trigger coverage for the domain, but a few common variations (e.g. 'training run', 'GPU server') are absent, so it sits at 'good keyword coverage, a few natural terms missing' rather than comprehensive.	4 / 5
Distinctiveness Conflict Risk	The niche is tightly scoped to RunPod GPU pods for explicitly chosen long-running research experiments with SSH access, yielding distinct triggers and minimal overlap with other skills.	5 / 5
	Total	17 / 20 Passed

Validation

100%

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 — 16 / 16 Passed

Validation for skill structure

No warnings or errors.

Repository: companion-inc/feynman
Path: skills/runpod-compute/SKILL.md
Commit: 95c51d9

Reviewed: about 9 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.