The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface through its automated discovery process. It is instructed to read various local project files and git logs to initialize the sales context.\n
Ingestion points: Local files including README, ICP/persona documents, playbooks, quota/comp/territory plans, agent-instruction files, and git logs (SKILL.md, Section 1).\n
Boundary markers: The instructions do not specify the use of delimiters or specific instructions to ignore embedded prompts within the ingested data.\n
Capability inventory: The skill has the capability to write and patch the sales-context.md file, influence routing decisions, and propose automated routines.\n
Sanitization: There is no evidence of sanitization or filtering applied to the content retrieved from external project files before it is processed.\n- [COMMAND_EXECUTION]: The skill relies on executing system commands to perform its discovery tasks.\n
Evidence: The instructions direct the agent to "read the recent log" of the repository's git history (SKILL.md, Section 1.2), which involves executing shell commands like git log.\n- [EXTERNAL_DOWNLOADS]: The skill references and recommends various sibling repositories and skills managed by the same author.\n
Evidence: The routing table and recommendation sections (SKILL.md, Section 3) point to resources such as mbfinotti/sales-skills, mbfinotti/revops-skills, and mbfinotti/partnerships-skills. These are vendor-owned resources.

sales-kickoff