The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill requires cloning a repository from an unverified third-party account (UrRhb/agentflow) during the installation process.
[REMOTE_CODE_EXECUTION]: Following the download, the user is instructed to install and execute shell scripts and AI worker processes directly from the unverified source.
[COMMAND_EXECUTION]: The skill establishes persistence by requiring the user to configure a crontab entry (*/15 * * * * ~/.claude/sdlc/agentflow-cron.sh) that executes the orchestrator script every 15 minutes.
[PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection because it processes untrusted data from external sources (Kanban board comments and SPEC.md) without documented sanitization.
Ingestion points: Kanban board task comments and SPEC.md file content (SKILL.md).
Boundary markers: None identified; the skill does not specify the use of delimiters or instructions to ignore embedded commands in ingested data.
Capability inventory: The skill can execute shell commands including npm test, tsc, eslint, and git revert, as well as dispatching autonomous worker agents.
Sanitization: No sanitization, escaping, or validation logic is described for the data processed from the Kanban board or specification files.

agentflow