The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses local shell commands to discover and process Dockerfiles. These commands are executed within the user's project environment to provide the skill's core functionality.
Evidence: SKILL.md defines a complex shell pipeline: bash -c 'find . \( -name '"'"'Dockerfile'"'"' -o -name '"'"'Dockerfile.*'"'"' -o -name '"'"'Containerfile'"'"' \) -type f -not -path '"'"'*/node_modules/*'"'"' -print0 | xargs -0 -r hadolint'.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from local Dockerfiles and interprets tool outputs, which could potentially contain instructions intended to mislead the agent.
Ingestion points: The skill reads and analyzes files matching the Dockerfile pattern within the local repository (referenced in SKILL.md Step 4 and toolchain.md).
Boundary markers: There are no specific delimiters defined to isolate the output of the hadolint tool from the agent's instructional context.
Capability inventory: The skill possesses file-writing capabilities for automatic fixes and the ability to execute shell commands like bash and hadolint as specified in the Execution Contract of SKILL.md.
Sanitization: While the command uses -print0 and xargs -0 to securely handle filenames, there is no sanitization or filtering of the content inside the Dockerfiles or the diagnostic messages returned by the linter.

codeguard-dockerfile