The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes several local Python scripts to handle the complex logic required for environment setup and model resolution.
python3 scripts/check-models-config.py: Resolves model aliases and reasoning efforts across multiple configuration layers.
python3 scripts/skill-listing.py: Manages how skills are listed in the Claude Code settings file to control invocation behavior.
python3 ../pstack-harness/scripts/subagents.py: Executes checks for persona readiness and registered subagents.
[INDIRECT_PROMPT_INJECTION]: The skill processes external configuration data that could theoretically contain adversarial instructions aimed at influencing the agent's setup process.
Ingestion points: The skill reads data from ~/.agents/pstack-models.md, .agents/pstack-models.md, ~/.claude/settings.json, and project-level .codex/config.toml files.
Boundary markers: No explicit boundary markers or 'ignore instructions' warnings are implemented when parsing the natural language mapping of roles to models.
Capability inventory: The skill has the ability to execute local scripts, run Git commands for repository management, and write or modify agent configuration and settings files.
Sanitization: A specialized linter (check-models-config.py) is used to validate the model configuration file against a predefined schema of roles and reasoning efforts before any changes are committed to disk.

setup-pstack