The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the wednesday-skills command-line tool to identify gaps in the codebase graph and perform analysis. These actions are performed using vendor-owned tools and are consistent with the onboarding workflow described in the skill metadata.
[PROMPT_INJECTION]: The skill reads content from project documentation (MASTER.md, PLAN.md) and git log to answer user questions, which introduces a surface for indirect prompt injection.
Ingestion points: MASTER.md, summaries.json, dep-graph.json, PLAN.md, and output from git log as specified in SKILL.md.
Boundary markers: Absent; there are no instructions to the agent to ignore or delimit potentially malicious content within the documentation files.
Capability inventory: Shell command execution via the Bash tool is restricted to wednesday-skills and git log as defined in the skill's permissions.
Sanitization: Absent; the skill does not specify any validation or filtering for the content read from the repository files.

onboard-dev-agent