This stack is for reviewing code, repositories, artifacts, or any surface where findings, bugs, and observations need to be captured and held without premature clustering, false generalization, or loss of evidence. The default failure mode in unstructured AI review sessions is that observations get collapsed into conclusions too early — verified states and hypothesized states get mixed, and the review produces a summary rather than a structured evidence record. This stack prevents that.Documentation Index
Fetch the complete documentation index at: https://mintlify.com/xxyoudeadpunkxx/ai-protocol-kit/llms.txt
Use this file to discover all available pages before exploring further.
Field Findings & Bugs Protocol v2
Capture findings as structured artifacts. Separate verified, inferred, and hypothetical states. Register observations without promoting them — a finding is recorded and held before any clustering, pattern-labeling, or resolution is attempted. This protocol controls the capture layer of the session.View protocol →
PHI-Lens Protocol v4.a (conditional)
Add this step only if constraints conflict and a flat compromise would be misleading or would hide the dominant force. The PHI-Lens protocol handles non-trivial constraint interactions — it makes asymmetry explicit rather than resolving it with a false middle. Do not add it as a default reviewer.View protocol →
The PHI-Lens step is conditional. Add it only when constraints actively interact and a straightforward resolution would hide which force is actually dominant. It is not a default second step for every review session — most captures do not need it. If findings are messy but not structurally conflicted, Step 1 alone is the right tool.
When to use this stack
- Code reviews where bugs, observations, and hypotheses need to stay separated rather than collapsed into a single summary
- Repository audits where the state of multiple files or components must be recorded without premature synthesis
- Bug capture sessions where evidence integrity matters more than a fast resolution narrative
- Any analytical work — reading artifacts, reviewing outputs, inspecting systems — where the findings must remain inspectable and promotable without loss of their original evidence basis