The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes the plannotator command with context-derived arguments using a bash shell. Evidence: Found in SKILL.md as plannotator annotate $ARGUMENTS.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection by acting upon feedback returned from tool outputs.
Ingestion points: Feedback content provided in the output of the plannotator annotate command as described in SKILL.md.
Boundary markers: Absent; the skill lacks instructions to delimit or ignore potentially malicious instructions within the feedback.
Capability inventory: Execution of shell commands via the plannotator tool and subsequent model interaction based on its output.
Sanitization: Absent; the skill instructs the agent to address feedback directly within the conversation without filtering.

plannotator-annotate