The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides and instructs the agent to execute a local bash script (scripts/plan-lint.sh) to validate generated implementation plans. This script uses standard tools like grep and awk for validation and is a core part of the skill's quality assurance process.
[PROMPT_INJECTION]: The skill exhibits a surface for indirect prompt injection because it ingests untrusted data (external specifications or requirements) to generate executable plans and performs subsequent validation steps.
Ingestion points: The "Pre-Writing" phase (SKILL.md) explicitly instructs the agent to read and study external specifications or requirements to gather context.
Boundary markers: The skill lacks explicit instructions or markers to delimit external content or to ignore potential embedded instructions within the input specifications.
Capability inventory: The skill possesses the ability to execute shell commands (via the bundled plan-lint.sh script) and write to the local file system (saving plans in the docs/superpowers/plans/ directory).
Sanitization: No explicit sanitization, filtering, or validation of the input specification content is defined in the instructions before the agent processes the data.

writing-plans-enhanced