skf-export-skill
Pass
Audited by Gen Agent Trust Hub on Oct 11, 2026
Risk Level: SAFECOMMAND_EXECUTIONDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [COMMAND_EXECUTION]: The workflow executes several local Python helper scripts (e.g., skf-manifest-ops.py, skf-rebuild-managed-sections.py) to handle manifest management, validation, and the rebuilding of managed sections in context files.
- [DYNAMIC_EXECUTION]: The skill supports a configurable on_complete hook and activation steps (activation_steps_prepend, activation_steps_append) defined in customize.toml, allowing for user-defined command execution.
- [INDIRECT_PROMPT_INJECTION]: The skill processes documentation and metadata from other skills to generate context snippets for IDE integration. Ingestion points: Reads metadata.json and SKILL.md of target skills in references/load-skill.md. Boundary markers: Employs and markers as defined in assets/managed-section-format.md. Capability inventory: Modifies system-level IDE configuration files such as CLAUDE.md and AGENTS.md in references/update-context.md. Sanitization: Uses stdin redirection in references/update-context.md §9 to prevent shell expansion of injected snippet content during the write process.
- [EXTERNAL_DOWNLOADS]: The skill documentation references the use of the 'skills' CLI tool via npx for publishing to and installing from the agentskills.io registry.
Audit Metadata