The Agent Skills Directory

[COMMAND_EXECUTION]: The skill includes a PowerShell script at scripts/bootstrap_eval_workspace.ps1 intended for setting up evaluation environments. The script automates the creation of directory structures and basic JSON metadata files using standard PowerShell cmdlets like New-Item and Set-Content. It does not perform network operations or access sensitive system configuration files.
[PROMPT_INJECTION]: The skill presents an indirect prompt injection surface (Category 8) as it is designed to ingest and interpret architecture and governance documentation from potentially untrusted project repositories.
Ingestion points: The skill reads BACKLOG.md, AGENTS.md, CHANGELOG.md, ADR files, and other architectural documents during the Inventory and Classification and Pattern Comparison stages (SKILL.md steps 4 and 5).
Boundary markers: Absent. The instructions do not specify delimiters or warnings to the agent to treat the contents of these repository files as untrusted data rather than instructions.
Capability inventory: The skill has the capability to write to CHANGELOG.md and BACKLOG.md to record findings and follow-up tasks (SKILL.md step 8).
Sanitization: The skill does not mention any sanitization or validation procedures for the data extracted from the repository files before it is processed by the agent or included in the final scan report.

ai-project-scan