The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze user-provided Yii2 codebases, creating a potential surface for indirect prompt injection where malicious instructions hidden in code comments or strings could influence agent behavior during extraction or semantic search phases.
Ingestion points: Path-to-project inputs in Phases 1, 4, and 5 allow external data into the agent's context.
Boundary markers: No specific boundary markers or instructions to treat codebase content as untrusted data are present.
Capability inventory: The skill uses the codecompass CLI for business logic extraction, semantic indexing, and search operations across local databases.
Sanitization: No explicit sanitization or filtering of the extracted codebase content is described in the workflow.
[COMMAND_EXECUTION]: The workflow involves running multiple shell commands, which is expected for a development tool but represents an execution surface.
Evidence: Instructions include running codecompass analyze:yii2, codecompass requirements:extract, codecompass batch:index, and codecompass search:semantic.
Local checks: The skill also uses curl to verify the schema of a local Weaviate instance on localhost:8081.
[EXTERNAL_DOWNLOADS]: The skill references the use of standard package managers to fetch external code.
Evidence: Phase 1.2 suggests running composer install if the vendor/ directory is missing, which downloads PHP dependencies from the public Packagist registry.

analyze-yii2-project