The Agent Skills Directory

[SAFE]: No malicious patterns or security risks were identified in the skill instructions or metadata.
[INDIRECT_PROMPT_INJECTION]: The skill ingests data from project files (SESSION_BRIDGE.md and HANDOFF.md) to facilitate task summarization. While specific boundary markers are absent, the instructions mandate sanitization through redaction and the skill contains no dangerous capabilities such as arbitrary command execution or network exfiltration, which minimizes the impact of processing untrusted file content.
[SAFE]: The skill includes security-best-practice instructions to proactively redact API keys, passwords, bearer tokens, and private credentials before summarizing data for a new session.

session-bridge