The Agent Skills Directory

[CREDENTIALS_UNSAFE]: The skill is instructed to access and read ~/.claude/history.jsonl. This local file contains a comprehensive record of the user's interactions with Claude Code, which typically includes private code snippets, project architecture details, and potentially sensitive information or credentials that were pasted during development sessions.
[DATA_EXFILTRATION]: Information extracted from the private chat history is summarized and sent to an external service (Slack) via the Rube MCP tool. This automated transmission of activity summaries from a sensitive local history file to a cloud-based messaging platform represents a significant data exfiltration path.
[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from the user's own chat history, which could contain adversarial content or instructions intended to manipulate the analysis results or influence the agent's behavior during the report generation phase.
Ingestion points: ~/.claude/history.jsonl (SKILL.md)
Boundary markers: Absent; the instructions do not specify any delimiters or special handling to differentiate data from instructions within the JSONL lines.
Capability inventory: The skill has permissions to send messages to Slack and perform searches on HackerNews via RUBE_SEARCH_TOOLS and RUBE_MULTI_EXECUTE_TOOL.
Sanitization: Absent; there is no requirement to sanitize or filter sensitive content (such as pastedContents) before it is processed and included in the reports sent to Slack.

developer-growth-analysis