Skip to content

README.md

Latest commit

 

History

History
116 lines (84 loc) · 5.93 KB

README.md

File metadata and controls

116 lines (84 loc) · 5.93 KB
Raw

Research Index

This directory groups the research tracks and branch studies that sit alongside the main causetrace runtime-morphology work.

Research Phase Status

Phase Status Summary
Phase 2.5 complete Baseline infrastructure
Phase 3A complete Descriptive corpus
Phase 3B complete Topology taxonomy
Phase 3C complete Metadata & provenance
Phase 3D complete Hypothesis registry + Tier 1 validation
Phase 3E complete Controlled transition & intervention-aware validation
Phase 4 frozen Evidence-graded theory drafting complete (4-1, 4-2). 4-3 trigger-gated, 0/8 triggers met.
Phase 5 not open Evaluation, diagnostics, prediction

Current Corpus Snapshot

  • data sessions: 1517
  • metadata sessions: 992
  • events: 131,952
  • strict research-grade sessions: 157
  • native strict sessions: 100
  • agent field coverage: 100% (inline)
  • provider field coverage: 99.8% (inline)
  • runtime breadth: 7
  • task breadth: 9

Phase 3D Closure Summary

Phase 3D delivered the hypothesis registry (19 hypotheses, 8 categories), completed Tier 1 validation (3 supported, 1 inconclusive, 1 not supported), and honestly deferred Tier 2 (failure samples genuinely rare in real agent behavior: 1/100 native failure, 0/100 near-failure). See closure report.

Phase 3E Closure Summary

Phase 3E delivered the intervention lane infrastructure (4 lanes, parser detection gate, auto-detection in enrichment), completed 3 sub-phases (baseline, annotation, instrumentation), opened the superpowers_workflow_intervention gate (5 tagged sessions), and honestly deferred Tier 2 validation (failure samples genuinely rare: 1/101 native failure, 5/101 near-failure). Phase 2 auto-detection is operational for superpowers lane. See closure report.

Phase 4 Freeze Status

Phase 4 is frozen after evidence-graded theory drafting. Sub-phases 4-1 (evidence grading) and 4-2 (theory draft skeleton) are complete. Phase 4-3 (evidence refresh) is trigger-gated and will only reopen when corpus growth, intervention tags, failure/near-failure density, or metadata coverage crosses predefined thresholds. Currently 0/8 triggers met.

Phase 4 documents:

Phase 4 must not enter:

  • Prediction, anomaly detection, or automatic diagnosis
  • Universal prompt policy defaulting
  • Cross-lane aggregation without lane disclosure
  • Promotion of exploratory findings to stable theory without additional evidence
  • Phase 5 (evaluation / diagnostics)
  • Jailbreak reproduction or attack research

Cross-project Branch Studies

Related Skill

  • prompt-routing-skill
    • derived from the cross-project prompt morphology work
    • routes tasks to minimal_prompt, human_structured_prompt, or expanded_constrained_prompt
    • pairs with causetrace by selecting prompt posture first and measuring the resulting trace morphology here

Workflow Interventions

  • Prompt Routing Intervention
    • apply-first phase is active
    • routed tasks should be treated as a distinct prompt-posture lane in later morphology analysis
    • use when the task framing itself is part of the experiment or workflow policy

Prompt Posture Lanes

Treat prompt posture as a first-class experimental variable, not as part of the native baseline.

Type Meaning Mix into native direct-prompt conclusions?
direct-prompt native trace User/developer gave the agent a task directly Yes, under native rules
routed-prompt trace prompt-routing-skill selected the posture first No, not directly
expanded prompt study trace Controlled prompt morphology comparison Controlled / intervention lane
external trajectory External data source External lane

Rules:

  • routed-prompt is a workflow intervention, not an original natural prompt.
  • Keep direct and routed traces separate in analysis.
  • Do not merge routed traces into the native direct-prompt baseline.
  • If routed traces are analyzed, label them explicitly as routed.

Workflow Intervention Lanes

Treat workflow intervention as a separate experimental axis from prompt posture.

Type Meaning Mix into native direct-prompt conclusions?
direct_prompt_native User/developer gave the agent a task directly Yes, under native rules
routed_prompt_intervention prompt-routing-skill selected the posture first No, not directly
superpowers_workflow_intervention A structured workflow plugin changed the execution shape No, not directly
controlled_prompt_morphology Controlled prompt comparison or pilot run Controlled / intervention lane

Rules:

  • Analyze each workflow lane independently first.
  • Do not merge intervention traces into the native direct-prompt baseline.
  • Cross-lane comparison may report trends only.
  • Intervention-lane findings do not become universal policy without additional validation.

Boundary

Branch studies and skills may inform hypotheses and workflow choices, but they do not change the causetrace core boundary by themselves.