sandbox-execution-fallback

Recover from execute_code_sandbox failures by writing Python scripts to files and executing via run_shell

Quality

66%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Critical

Do not install without reviewing

Fix and improve this skill with Tessl

tessl review fix ./benchmarks/gdpval/skills/sandbox-execution-fallback/SKILL.md

Quality

Content

75%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.

The body delivers highly actionable, copy-paste-ready guidance with a clear sequenced workflow and verification, though it carries some padding in supplementary sections that explain concepts Claude already knows.

Suggestions

Remove or shrink the 'Why This Works' comparison table; Claude already understands sandbox-vs-shell differences in environment, file I/O, and timeouts.

Drop generic best-practice items that restate basic programming knowledge (e.g., 'Include try/except blocks', 'Use print statements for debugging') and keep only skill-specific guidance.

Merge 'Common Use Cases' into 'When to Use' to eliminate redundancy, and add an explicit feedback step for when Step 4 verification fails (not just when run_shell fails).

Dimension	Reasoning	Score
Conciseness	The core procedure and complete example are efficient, but supplementary sections (the 'Why This Works' comparison table explaining sandbox-vs-shell basics, generic best-practice advice like 'Include try/except blocks', and a 'Common Use Cases' list that duplicates 'When to Use') add unnecessary explanation Claude already knows.	3 / 5
Actionability	Guidance is fully executable and copy-paste ready: concrete write_file(path=..., content=...) and run_shell(command='python3 ...') calls plus a complete pandas/openpyxl example that covers the common case end-to-end.	5 / 5
Workflow Clarity	A clear 4-step sequence (capture, write, execute, verify) includes an explicit Step 4 'Verify Output' checkpoint and a Troubleshooting feedback loop for run_shell failures, but lacks an integrated 'if verification fails, re-run' loop tied to output correctness.	4 / 5
Progressive Disclosure	Content is self-contained with no bundle files and is organized into clear, well-labeled sections with no nested references; it scores 4 rather than 5 because the skill exceeds 50 lines and supplementary sections (Why This Works, Common Use Cases) could be trimmed or split.	4 / 5
	Total	16 / 20 Passed

Description

58%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.

The description is specific and highly distinctive with explicit what/when content, but relies on technical jargon rather than natural user trigger phrases.

Suggestions

Add a natural-language 'Use when...' clause with user-facing trigger phrases (e.g., 'Use when sandbox code execution fails, times out, or lacks required libraries').

Soften the jargon by including recognizable synonyms users actually say (e.g., 'code execution failed', 'sandbox timed out') alongside the tool names.

Broaden the action list slightly (e.g., 'install missing packages, then re-run') to push specificity toward comprehensive coverage.

Dimension	Reasoning	Score
Specificity	Names the domain (sandbox failure recovery) and two concrete actions ('writing Python scripts to files', 'executing via run_shell'), matching the anchor for 1-2 concrete actions without being comprehensive enough for 4.	3 / 5
Completeness	Both what (write scripts to files, execute via run_shell) and when (execute_code_sandbox failures) are explicit, with the trigger embedded in 'Recover from execute_code_sandbox failures' as equivalent trigger guidance, though a standalone 'Use when' clause would be clearer.	4 / 5
Trigger Term Quality	The terms ('execute_code_sandbox', 'run_shell', 'Python scripts') are mostly technical jargon rather than natural phrases a user would say; 'Python scripts' is the only mildly natural term, so it sits just above the pure-jargon anchor.	2 / 5
Distinctiveness Conflict Risk	The trigger 'execute_code_sandbox failures' defines a narrow, distinct niche with virtually no overlap risk against other skills.	5 / 5
	Total	14 / 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: HKUDS/OpenSpace
Path: benchmarks/gdpval/skills/sandbox-execution-fallback/SKILL.md
Commit: 3827781

Reviewed: 15 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.