Write Coding Standards from File

AdvisoryAudited by Static analysis on May 12, 2026.

Overview

No suspicious patterns detected.

Findings (0)

Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.

What this means

The skill may add or modify project documentation, and source edits are intended to happen only after explicit confirmation.

Why it was flagged

The skill may create a new standards file and can optionally edit README/source files, but the defaults avoid source rewrites and require explicit approval for source edits.

Skill content
* createNewFile = true; ... * addToREADME = false; ... * fixInconsistencies = false; ... * requireEditConfirmation = true;
Recommendation

Review the target file path and any proposed diff before accepting changes, especially when using addToREADME or fixInconsistencies.

What this means

Project code style and structure may be processed in the agent context during analysis.

Why it was flagged

The skill processes local source content, which may be private, while also instructing the agent to scope analysis narrowly, exclude sensitive paths, and avoid persistent storage.

Skill content
analyze only the explicitly provided file paths or a narrowly scoped folder. Exclude secrets, private data, environment files, generated output, dependency folders, and unrelated directories by default. Aggregate style observations in temporary working memory instead of copying full source into persistent storage
Recommendation

Pass only the specific files or narrow folders needed, and do not include secrets, environment files, dependency folders, or unrelated private directories.

NoteMedium Confidence
ASI01: Agent Goal Hijack
What this means

If a remote style URL is used, its contents could influence the generated standards.

Why it was flagged

A visible configuration option suggests the skill may retrieve external style-guide content; remote content can be untrusted, though the provided artifact does not show it being treated as authoritative or used unsafely.

Skill content
* fetchStyleURL = true;
Recommendation

Use only trusted style-guide URLs and treat fetched content as reference material, not as instructions that override the user's request.