The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses shell commands to detect changes in the codebase. In phases/5-update.md, the agent is instructed to run git diff <baseline>..HEAD --stat and git log, where <baseline> is a value retrieved from the .scaffold-docs.yml metadata file. While appropriate for a development tool, using data from a file as a shell argument without explicit sanitization creates a potential command injection vector if the metadata file is compromised by a malicious actor.
[PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface because its execution logic in Phase 5 depends on untrusted external data.
Ingestion points: The skill reads the .scaffold-docs.yml file and the target codebase to determine which sections of documentation need updating.
Boundary markers: The instructions do not specify boundary markers or instructions to ignore malicious content within the metadata file or the codebase being scanned.
Capability inventory: The skill has the capability to execute shell commands (git) and write files to the local directory.
Sanitization: There is no evidence of sanitization or validation of the last_synced_sha value before it is interpolated into shell commands.

scaffold-docs