The Agent Skills Directory

[COMMAND_EXECUTION]: The skill autonomously executes shell commands to perform code validation and testing across multiple tech stacks, including php -l, node --check, kotlinc, npm test, and gradlew test.
[COMMAND_EXECUTION]: The executor performs automated Git operations, including git add, git commit, and git push, to sync completed implementation phases with the remote repository.
[PROMPT_INJECTION]: The instructions command the agent to adopt a persona of 'full executive authority' and explicitly direct it to 'not ask for permission' and 'not interrupt' the user, which reduces the opportunity for human oversight during critical implementation steps.
[PROMPT_INJECTION]: The skill is designed to ingest and act upon untrusted data from external files, creating a surface for indirect prompt injection.
Ingestion points: Tasks and requirements are read from markdown files in the docs/plans/ directory (e.g., SKILL.md Step 0).
Boundary markers: There are no explicit markers or warnings used to separate the external plan data from the core system instructions.
Capability inventory: The agent has the capability to write files, execute arbitrary test scripts/commands, and push data to a remote repository.
Sanitization: The skill employs a '5-Layer Validation Stack' (Layer 4) that specifically checks for security vulnerabilities like SQL injection and XSS in the generated output.

plan-implementation