The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: Automated security scanners (URLite) have flagged the documentation URL https://jeffallan.github.io/claude-skills/skills/specialized/legacy-modernizer/ as malicious, resulting in its inclusion on a blacklist.
[METADATA_POISONING]: The skill's main configuration file SKILL.md has been flagged by reputation scanners as malicious (FileRepMalware), indicating that the skill itself contains known malicious patterns or originates from an untrusted source.
[COMMAND_EXECUTION]: The skill includes code in references/system-assessment.md that utilizes subprocess.run to execute shell commands (git log) on the project directory. Executing arbitrary commands on external repositories without rigorous input validation or sandboxing presents a significant risk.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest large amounts of untrusted external data (the user's source code) for analysis, which creates a substantial attack surface for indirect prompt injection.
Ingestion points: The LegacyCodeAnalyzer class in references/system-assessment.md recursively reads all .py files within a project directory using open().read().
Boundary markers: The skill lacks any delimiters or specific instructions to the agent to treat file contents as untrusted data or to ignore instructions embedded within the codebase.
Capability inventory: The skill has powerful capabilities, including file system access and the ability to execute shell commands via subprocess.
Sanitization: There is no evidence of content sanitization or safety checks before the ingested code is processed by the agent's logic.

legacy-modernizer