The Agent Skills Directory

[COMMAND_EXECUTION]: The skill relies on shell scripts (fetch-docs.sh, fetch-raw.sh, extract-code-blocks.sh, extract-notes.sh, and extract-signatures.sh) to orchestrate documentation retrieval and perform text processing. These scripts are executed directly by the agent environment.
[EXTERNAL_DOWNLOADS]: The skill uses npx -y @upstash/context7-mcp to download and execute the Context7 MCP server from the npm registry at runtime. While Upstash is a well-known service, the use of npx -y introduces a dependency on external code that is resolved and executed during skill operation.
[DYNAMIC_EXECUTION]: The scripts/mcp-client.py utility uses subprocess.Popen with shell=True to spawn the MCP server process. Although the command string is currently hardcoded in the shell scripts, this execution pattern increases the risk of command injection if user-controlled input were to reach the command string.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and summarize external documentation, which constitutes a potential attack surface for indirect prompt injection.
Ingestion points: Documentation content is fetched from external sources via the Context7 MCP server.
Boundary markers: The skill does not implement explicit boundary markers or instructions to the agent to treat fetched documentation as untrusted data.
Capability inventory: The skill has the ability to execute shell scripts and perform subprocess calls.
Sanitization: While the skill filters the documentation structurally (extracting code blocks, signatures, and notes using awk and grep) to save tokens, it does not perform content-level sanitization to identify or strip potential prompt injection payloads embedded in the documentation.

context7-efficient