The Agent Skills Directory

[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests and processes untrusted data from the codebase being profiled.
Ingestion points: Phase 3 reads codebase source files (e.g., _process, Update, Tick functions), and Phase 2 reads performance targets from design documents or AGENTS.md.
Boundary markers: Absent; there are no instructions to use delimiters or warnings to ignore instructions embedded in the analyzed code or documentation.
Capability inventory: The skill has access to the Bash, Read, Glob, and Grep tools, which allow for broad file system interaction and command execution.
Sanitization: Absent; the skill does not specify any validation, filtering, or escaping of the content it reads before processing it.
[COMMAND_EXECUTION]: The skill utilizes the Bash tool to perform technical analysis of the project's source code.
The instructions in Phase 3 involve scanning the codebase for specific patterns (loops, string operations, physics queries), which are typically implemented via shell commands.

perf-profile