The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill's 'Audit mode' employs multi-agent orchestration to review external Go source code for observability signals.
Ingestion points: The signal-specific sub-agents (metrics, logging, tracing, profiling, RUM) ingest and process content from Go source files (**/*.go) using the Agent tool.
Boundary markers: The orchestration instructions lack explicit delimiters or instructions to ignore embedded commands within the audited codebase, which could allow malicious code to influence the sub-agents' analysis.
Capability inventory: The skill utilizes significant capabilities including Bash commands, file operations, and sub-agent creation, which are part of the processing pipeline for external data.
Sanitization: The skill does not describe any sanitization or validation steps for the code content before it is processed by the auditing sub-agents.
[COMMAND_EXECUTION]: Shell command execution is strictly controlled through tool-prefix constraints in the configuration.
The Bash tool is restricted to well-defined prefixes for Go tooling (go:*), linting (golangci-lint:*), and version control (git:*), effectively limiting the agent's execution scope to the project's development workflow.
[EXTERNAL_DOWNLOADS]: Recommends the integration of various observability libraries and SDKs from established sources.
Mentions vendor-specific tools from the author's ecosystem (samber/slog-*, samber/oops) and well-known services like PostHog, Segment, and Pyroscope for monitoring and analytics.
These references target official repositories and documentation for legitimate development purposes.

golang-observability