The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructions mandate the use of the Bash tool to execute a local Python script: python3 "${CLAUDE_PLUGIN_ROOT}/tools/saga_driver.py" --layer 06_SPEC --allow-skip-permissions.
[PRIVILEGE_ESCALATION]: The workflow explicitly utilizes the --allow-skip-permissions flag. The skill's documentation confirms this is intended to bypass standard permission prompts for file writes, enabling "unattended autopilot" which removes user oversight from potentially dangerous file system operations.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process untrusted external data which could contain malicious instructions.
Ingestion points: The skill reads BDD/ADR documents, implementation plans (IPLAN), and free-text user prompts as primary sources for specification generation.
Boundary markers: No explicit boundary markers or instructions to ignore embedded commands within the source documents are provided in the workflow.
Capability inventory: The skill has access to the Bash tool and file system write capabilities via the saga_driver.py script and the doc-spec-fixer dependency.
Sanitization: There is no evidence of sanitization or validation performed on the content of the upstream artifacts before they are interpolated into the generation and audit logic.

doc-spec-autopilot