The Agent Skills Directory

[DYNAMIC_CONTEXT_INJECTION]: The skill uses the dynamic context syntax to execute a shell command at load time. * Evidence: !ls .github/workflows/*.yml .gitlab-ci.yml Jenkinsfile .circleci/config.yml 2>/dev/null || echo 'No CI/CD config found'`` in SKILL.md. * Analysis: The command is a benign file discovery operation used to inform the agent about the current project state. it does not perform network operations or access sensitive credentials.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted configuration files from the user's project environment to generate CI/CD pipelines, creating a potential vector for malicious instructions to be ingested. * Ingestion points: SKILL.md instructions specify scanning the project for existing CI/CD configuration files. * Boundary markers: Absent. there are no explicit instructions for the agent to ignore embedded instructions within the files it reads. * Capability inventory: The skill has access to Read, Write, Edit, Grep, Glob, and Bash tools. * Sanitization: Absent. the skill relies on the agent's default processing logic without explicit sanitization steps for the ingested data.

building-cicd-pipelines