debug

Investigate stuck runs and execution failures by tracing Symphony and Codex logs with issue/session identifiers; use when runs stall, retry repeatedly, or fail unexpectedly.

1.00x

Quality

92%

Does it follow best practices?

Impact

83%

1.00x

Average score across 3 eval scenarios

Securityby

Passed

No findings from the security scan

Quality

Content

85%Weight 40%Scale 1-3

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

The body is highly actionable with concrete rg commands and a clear sequenced debugging workflow, and is well-organized with no nested references. Its main weakness is redundancy across the three tracing sections that inflates token count.

Suggestions

Merge the overlapping tracing guidance in 'Quick Triage', 'Investigation Flow', and 'Reading Codex Session Logs' into a single canonical session-tracing section referenced by the others, to remove duplicated steps.

Move the failure-classification taxonomy (stall loop / app-server startup / turn failure / worker crash) into one place instead of restating it across sections.

Consider extracting the full rg command block into a references/ script so the main body keeps only the highest-leverage commands.

Dimension	Reasoning	Score
Conciseness	The body is lean with no concept re-explanation, but Quick Triage, Investigation Flow, and Reading Codex Session Logs each restate the same session_id tracing and failure-classification steps, which could be consolidated.	2 / 3
Actionability	Provides fully executable, copy-paste-ready rg commands with concrete log line patterns (e.g. 'session_id=[^ ;]+', 'Issue stalled\|...turn_timeout\|turn_failed').	3 / 3
Workflow Clarity	Multi-step processes (Quick Triage, Investigation Flow) are clearly sequenced with an explicit 'Validate scope' checkpoint and evidence-capture step; this is read-only log analysis so no destructive validate-fix-retry loop is required.	3 / 3
Progressive Disclosure	No bundle files exist, so the self-contained SKILL.md with well-organized sections (Goals, Log Sources, Correlation Keys, Commands, flows, Notes) and one-level doc reference satisfies progressive disclosure for a simple skill.	3 / 3
	Total	11 / 12 Passed

Description

100%Weight 40%Scale 1-3

Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.

The description is concise, third-person, and answers both what and when with concrete, domain-specific actions and natural trigger terms. It is a strong, low-conflict trigger description.

Dimension	Reasoning	Score
Specificity	Names multiple concrete actions ('Investigate stuck runs and execution failures by tracing Symphony and Codex logs with issue/session identifiers') tied to a specific domain, rather than vague language.	3 / 3
Completeness	Explicitly states both what it does (investigate/trace logs with identifiers) and when to use it via an explicit 'use when' trigger.	3 / 3
Trigger Term Quality	The 'use when runs stall, retry repeatedly, or fail unexpectedly' clause covers natural terms a user would actually say when needing this skill.	3 / 3
Distinctiveness Conflict Risk	The Symphony/Codex log + issue/session identifier niche is highly specific and unlikely to trigger for unrelated skills.	3 / 3
	Total	12 / 12 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: openai/symphony
Path: .codex/skills/debug/SKILL.md
Commit: 1f3219b

Reviewed: 12 days 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.