skill-store
Warn
Audited by Gen Agent Trust Hub on May 16, 2026
Risk Level: MEDIUMEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONREMOTE_CODE_EXECUTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill's installation process requires downloading the 'agentcli-helpers' package from PyPI or directly from a GitHub repository ('github.com/lirrensi/agent-cli-helpers').
- [COMMAND_EXECUTION]: The skill utilizes 'npx skills' to add external content to the local store. This involves downloading and executing packages from the npm registry at runtime.
- [REMOTE_CODE_EXECUTION]: The 'skill-store sync' command is designed to automatically extract zip-compressed '.skill' bundles within the local store directory. This represents a dynamic execution pattern where external code is processed and integrated into the agent's environment.
- [DATA_EXFILTRATION]: A surface for indirect prompt injection exists as the skill is specifically designed to ingest untrusted data from external sources (community packages and zip bundles) and load their content into the active agent context.
- Ingestion points: External skills are added via 'npx skills add' or manually placed '.skill' zip files, as documented in 'references/npx-skills.md' and 'references/importing.md'.
- Boundary markers: No specific boundary markers or 'ignore' instructions are provided to the agent when loading these external skill files.
- Capability inventory: The skill possesses file system access (read/write/zip extraction) and shell execution capabilities via the 'skill-store' and 'npx' CLIs.
- Sanitization: There is no evidence of content sanitization or validation of the ingested skill files before they are loaded into the agent's context.
Audit Metadata