skillsmp-search

Warn

Audited by Gen Agent Trust Hub on Mar 1, 2026

Risk Level: MEDIUMEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill is designed to download SKILL.md files from arbitrary GitHub repositories identified via the SkillsMP API.
  • These downloaded files are saved locally and act as executable instructions for the agent, creating a risk if the source repository is malicious.
  • [COMMAND_EXECUTION]: Shell commands are used for networking and local file access.
  • The skill utilizes curl to communicate with the SkillsMP API and to fetch remote files from raw.githubusercontent.com.
  • It executes cat to retrieve the stored API key from the local filesystem for authorization.
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection through community-contributed content.
  • Ingestion points: External SKILL.md files are downloaded and processed during the comparison and installation workflows.
  • Boundary markers: There are no explicit delimiters or instructions to ignore embedded commands within the ingested skill files.
  • Capability inventory: The skill possesses the ability to execute network requests via curl, read local files, and write new skill files to the filesystem.
  • Sanitization: The process performs basic frontmatter validation but does not include an audit or sanitization of the instructional content within the downloaded skills.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Mar 1, 2026, 01:33 PM