The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes various shell commands for project discovery and management.
It resolves the project root using git rev-parse --show-toplevel or pwd -P.
The validation setting flow executes a user-provided command string in the shell to verify its success before saving it to configuration files.
It performs file system inspections using ls, awk, sed, cat, and mkdir across the project structure.
[DYNAMIC_EXECUTION]: The bundled scripts/tickets.sh script dynamically executes a large block of Python code using a shell heredoc (`python3
"$@" <<'PY' ...`). This script is used to bridge interactions with external issue tracker APIs.
[DATA_EXFILTRATION]: The skill interacts with external services and explicitly configures data flow off-machine.
The peer reviewer setting allows the user to configure external CLI tools (e.g., codex, gemini, grok) to receive code diffs and briefs for review purposes.
The scripts/tickets.sh script sends issue data, labels, and comments to GitHub, Linear, and Jira APIs using authenticated network requests.
[INDIRECT_PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection through its ingestion of project configuration files.
Ingestion points: Reads CLAUDE.md, AGENTS.md, docs/agents/issue-tracker.md, and docs/agents/triage-labels.md to extract current settings and personas.
Boundary markers: Relies on markdown headers (## Agent skills) and specific line prefixes (Persona:, Style:) as delimiters.
Capability inventory: Includes shell command execution (bash, git, gh), Python execution, file writing, and network operations via issue tracker integrations.
Sanitization: Uses pattern matching (sed/awk) to extract configuration lines; however, it lacks strict validation or escaping for the values assigned to persona and style instructions beyond keyword matching for its own reference files.

attune