The Agent Skills Directory

[COMMAND_EXECUTION]: The skill invokes a local shell script scripts/install-explorer.sh during its setup phase to configure subagent definitions in the project environment. Additionally, subagents are permitted to use read-only Bash utilities (such as grep, find, and rg) to facilitate analysis of the targeted source code.
[EXTERNAL_DOWNLOADS]: The skill employs WebFetch and WebSearch capabilities to retrieve and process information from remote URLs as part of its multi-area research workflow.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it is designed to ingest and process untrusted data from external sources (codebases and web content).
Ingestion points: External data enters the agent context through Read, WebFetch, and WebSearch tools used by subagents (documented in SKILL.md and assets/explorer-agent.md).
Boundary markers: While the subagent is given a specific scope, the skill does not implement explicit structural delimiters or "ignore embedded instructions" markers for the ingested content.
Capability inventory: Subagents are granted restricted Write access to specific paths and read-only Bash access (defined in references/dispatch-rules.md).
Sanitization: The skill does not perform sanitization, escaping, or validation of the content retrieved from external sources before presenting it to the language model.

agent-exploration