The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill recommends an installation method for 'Continuous Claude' that pipes a remote shell script directly to bash (curl -fsSL https://raw.githubusercontent.com/AnandChowdhary/continuous-claude/HEAD/install.sh | bash). This allows arbitrary code execution from a source outside the trusted list.
[COMMAND_EXECUTION]: The skill's primary purpose is to orchestrate autonomous shell command execution through patterns like 'Sequential Pipelines'. This creates a significant attack surface if the agent is instructed to run unvalidated commands generated from external inputs.
[EXTERNAL_DOWNLOADS]: The skill references and downloads implementation scripts and configurations from third-party GitHub repositories (AnandChowdhary/continuous-claude) that are not recognized as trusted organizations or well-known services.
[PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection (Category 8) because it ingests untrusted data from multiple sources.
Ingestion points: The 'Infinite Agentic Loop' reads specification files (specs/component-spec.md), and the 'Continuous Claude PR Loop' reads PR descriptions and CI logs via gh run view.
Boundary markers: No specific boundary markers or instructions to ignore embedded commands are present in the provided templates.
Capability inventory: The skill possesses extensive capabilities including file system writes, network operations via curl, and shell command execution via claude -p and bash.
Sanitization: There is no evidence of sanitization or validation of the external content before it is processed by the agent.

autonomous-loops