ai-collab-protocols

To maintain AI workflow continuity across sessions, enforce stable URL-based entity references and use durable PR-comment threads as session memory. This prevents context loss by anchoring task planning to persistent identifiers rather than ambiguous labels.

Collaboration protocols for multi-step Claude workflows involve standardizing entity signals with stable URLs or resource identifiers. This process avoids pronoun ambiguity and emits concrete handle suggestions to ensure durable tracking throughout task execution.

Standardize entity references in AI task planning by converting informal labels into stable, URL-based handles. This requires strict reference handling for entities like PRs and bugs, ensuring continuous and unambiguous tracking across sessions.

Project rule files like AGENTS.md, CLAUDE.md, or .clinerules are detected to guide AI collaboration. If absent, the protocol prompts you to author them, ensuring consistent workflow rules and stable entity handling across sessions.

Pronouns should not be used to reference bugs or PRs in session memory due to ambiguity. Instead, emit concrete handle suggestions using stable URL-based identifiers to maintain strict entity reference continuity across sessions.

The best way to surface stable identifiers for AI workflows is to enforce URL-as-entity-reference for all mentioned entities. This converts unclear labels into durable handles, maintaining continuity across multi-step sessions.