The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill requires the installation of acpx from the NPM registry (npm install -g acpx@latest). This is an external dependency from a source not identified as a trusted organization or well-known service.
[REMOTE_CODE_EXECUTION]: The skill uses npx acpx@latest to execute remote code during runtime and includes a command to install further instructions from a remote source via npx acpx@latest --skill install acpx.
[COMMAND_EXECUTION]: The skill uses the acpx bridge to execute complex shell commands and delegate development tasks like fixing tests, refactoring modules, and running PR reviews.
[PRIVILEGE_ESCALATION]: The instructions explicitly recommend using the --approve-all flag. This flag is designed to auto-approve all tool execution requests, including destructive file writes and shell commands, effectively bypassing the agent platform's safety guardrails and human-in-the-loop requirements.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze external codebases and pull requests, creating a surface for indirect prompt injection given its high-privilege capabilities.
Ingestion points: Source code files, repository purpose summaries, and pull request changesets referenced in SKILL.md.
Boundary markers: Absent; the skill does not define delimiters or specific instructions to treat code data as untrusted content.
Capability inventory: The skill possesses full filesystem write access and shell execution capabilities through the acpx tool.
Sanitization: Absent; there are no instructions to sanitize, escape, or validate content retrieved from external repositories before processing.

qwen-code-claw