The Agent Skills Directory

[SAFE]: The skill documents a standard methodology for performance tuning, including the use of well-known diagnostic tools such as perf, py-spy, strace, and EXPLAIN ANALYZE. These tools are appropriate for the skill's intended purpose of backend bottleneck localization.
[PROMPT_INJECTION]: Analysis of the instructions shows no attempts to bypass safety filters, override agent constraints, or extract system prompts.
[DATA_EXFILTRATION]: No evidence of unauthorized data access or transmission was found. The skill references well-known monitoring and observability platforms like Prometheus, Sentry, and Grafana for legitimate telemetry purposes.
[COMMAND_EXECUTION]: The skill utilizes Bash and various CLI profiling tools. These operations are restricted to the task of performance measurement and do not exhibit malicious command patterns or persistence mechanisms.
[INDIRECT_PROMPT_INJECTION]: The skill possesses an attack surface for indirect prompt injection as it processes external codebases and profiling logs.
Ingestion points: Source code accessed via lsp MCP, profiling tool output, and database query logs.
Boundary markers: Absent.
Capability inventory: Bash (command execution), Edit (file modification), Read (file access).
Sanitization: Absent.

performance-optimization