The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes various TypeScript project files which could contain malicious instructions if the project files are from an untrusted source.
Ingestion points: Project configuration files such as package.json, tsconfig.json, biome.json, and pnpm-workspace.yaml are read to determine project state and runtime targets.
Boundary markers: Absent; the skill does not specify using delimiters or explicit warnings to ignore instructions embedded in the ingested data.
Capability inventory: The skill uses tools like pnpm, bash, tsc, and biome to write files and execute commands across the repository.
Sanitization: Absent; there is no evidence of sanitization for content read from the project configuration files.
[COMMAND_EXECUTION]: The skill instructs the agent to execute shell commands using pnpm, tsc, biome, and other CLI tools like npm-check-updates and depcheck for project management and validation tasks.
[DYNAMIC_EXECUTION]: The skill generates and configures shell scripts for agent hooks located at .claude/hooks/ts.sh and .github/hooks/ts.sh. These scripts are executed by the agent platform to perform automated formatting and type-checking upon file edits.

typescript