code-exec-fallback

Fallback pattern for executing Python code when execute_code_sandbox fails

Quality

50%

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

Fix and improve this skill with Tessl

tessl review fix ./benchmarks/gdpval/skills/code-exec-fallback/SKILL.md

Quality

Content

57%

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

The skill is well-organized and lean for a simple single-purpose task, but its step instructions lean on placeholders rather than fully executable commands and it lacks validation checkpoints around the run/cleanup loop. Verbosity in the tips and 'when this helps' sections slightly undercuts token efficiency.

Suggestions

Replace the prose-in-code-fence step instructions with concrete, copy-paste-ready write_file/run_shell invocations including a timeout value.

Add a verification checkpoint: check run_shell exit code and stderr before deleting the temp script, and only clean up on success.

Trim 'Tips for Success' and 'When This Helps' to the few items Claude would not already infer (e.g. the 2-attempt threshold, shell-environment dependency differences).

Dimension	Reasoning	Score
Conciseness	The body is mostly efficient with executable code, but sections like 'Tips for Success' and 'When This Helps' restate things Claude already knows (imports, absolute paths, JSON output) and pad the token budget somewhat.	2 / 3
Actionability	It provides a concrete worked example with write_file/run_shell calls, but Step 1/2/3 use prose placeholders inside code fences rather than truly executable commands, leaving the core steps as guided pseudocode.	2 / 3
Workflow Clarity	The write -> execute -> clean-up sequence is present, but there are no validation or verification checkpoints (e.g. check run_shell exit status / stderr before cleaning up), and cleanup of a temp file is a mildly risky operation executed without confirmation.	2 / 3
Progressive Disclosure	For a short single-purpose skill with no bundle files present, the content is well-organized into clear sections (When to Use, The Pattern, Step-by-Step, Example, Tips) at one level, which is appropriate.	3 / 3
	Total	9 / 12 Passed

Description

42%

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 specific about the concrete actions involved but lacks any explicit trigger guidance for when to use the skill, which is its main weakness. Trigger-term coverage is moderate and it carries some conflict risk with general code-execution skills.

Suggestions

Add an explicit 'Use when...' trigger clause, e.g. 'Use when execute_code_sandbox fails after repeated attempts due to timeouts, missing dependencies, or sandbox restrictions.'

Include more natural user-facing trigger terms such as 'run Python code', 'code won't run', or 'sandbox error' so users' phrasings match.

Tighten distinctiveness by scoping to the failure path, e.g. 'Fallback only — not for routine code execution.'

Dimension	Reasoning	Score
Specificity	Names concrete actions: "executing Python code", "write_file", "run_shell", and "python3 script.py" — multiple specific, actionable operations rather than abstract language.	3 / 3
Completeness	It states what the skill does (fallback for executing Python code) but provides no explicit "Use when..." clause or trigger guidance for when to invoke it, which caps completeness at 2 per the guidelines — and here even the 'when' is only weakly implied.	1 / 3
Trigger Term Quality	Contains some relevant terms ("execute_code_sandbox fails", "Python code", "sandbox") but misses common user-facing variations a person would actually say, such as "run code" or "code won't run".	2 / 3
Distinctiveness Conflict Risk	Tied to a specific failure mode of execute_code_sandbox giving it a niche, but the language overlaps with general code-execution skills and could trigger for ordinary run-code requests.	2 / 3
	Total	8 / 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: HKUDS/OpenSpace
Path: benchmarks/gdpval/skills/code-exec-fallback/SKILL.md
Commit: c5a9c4b

Reviewed: 2 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.