The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill includes functionality to download additional agent skills from remote repositories.\n
Evidence: Section 2.4 in SKILL.md contains instructions to use curl to fetch SKILL.md files from URLs defined in a local catalog.\n
Source: The target URLs point to official GitHub organizations for Firebase and the skill's author (gemini-cli-extensions), which are recognized as trusted or well-known entities.\n- [COMMAND_EXECUTION]: The skill uses shell commands for project orchestration and environment setup.\n
Evidence: It executes mkdir to create directories, curl to download files, and git to commit changes to the repository.\n
Context: These operations are consistent with the skill's stated purpose of managing a development workflow.\n- [PROMPT_INJECTION]: The skill presents an attack surface for indirect prompt injection by ingesting external data into its reasoning process.\n
Ingestion points: Reads foundational project files (product.md, tech-stack.md, workflow.md) and track-specific artifacts (spec.md, plan.md).\n
Boundary markers: Absent; there are no instructions to the agent to treat data from these files as untrusted or to ignore instructions within them.\n
Capability inventory: The agent can execute shell commands, write to the file system, and modify project history via Git.\n
Sanitization: Absent; the skill interpolates data from these files directly into the context used for generating new technical documents.

conductor-new-track