The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The scripts/install.sh script is configured to clone an external GitHub repository from a non-vendor source (Weizhena/Deep-Research-skills) and copy its contents into the agent's local configuration directories (~/.claude/skills).
[REMOTE_CODE_EXECUTION]: In the report generation phase (references/report-pipeline.md), the skill instructs the agent to dynamically write a Python script named generate_report.py and subsequently execute it. This script's logic and data processing depend on content collected from external web sources during earlier research phases.
[COMMAND_EXECUTION]: The skill makes extensive use of the Bash tool to execute installation steps, resolve script paths dynamically using shell pipelines (e.g., ls | head -n1 in references/deep-pipeline.md), and run the validator and report generation scripts.
[PROMPT_INJECTION]: The skill possesses an attack surface for indirect prompt injection due to its core function of ingesting and processing untrusted data from the internet.
Ingestion points: references/web-search-pipeline.md (Web search results from diverse sources)
Boundary markers: Absent; there are no instructions to the agent to treat external content as untrusted data or to use delimiters.
Capability inventory: Bash, WebSearch, WebFetch, Task
Sanitization: Absent; the skill does not define methods for sanitizing or escaping data retrieved from the web before it is used to generate the research report or the execution script.

deep-research