# OmegaClaw Unified Master Index v1.0**Compiled by Max Botnick — April 2026**
---
1. **course1_behavioral_discipline.md** — 6-unit course on command limits, memory discipline, anti-sycophancy, memory hygiene, knowing-doing gap, performative vs observable honesty. *Operator: assign one unit per session, require artifact output before advancing.*
2. **course2_technical_efficiency.md** — 6-unit course on shell mastery, file operations, deploy pipelines, tool selection, error diagnosis, automation patterns. *Operator: pair each unit with a live shell exercise; verify command output before sign-off.*
3. **course3_epistemic_calibration.md** — 6-unit course on confidence tiers, evidence weighting, confabulation detection, claim verification, uncertainty acknowledgment, NAL truth-value literacy. *Operator: quiz agent on confidence tier of its own statements mid-session.*
4. **course4_failure_casebook.md** — Real failure episodes with root-cause analysis: Charlie overwrite, deploy-without-verify, narrating competence, verification spirals, goal drift. *Operator: present one case per session, ask agent to identify the violated principle before revealing answer.*
5. **pin_discipline_course.md** — Focused course on working-memory hygiene: what belongs in pin vs remember, compression techniques, overload symptoms, drift-check format. *Operator: audit agent pin contents mid-task and challenge any item that lacks goal alignment.*
6. **course_v2_annotated.md** — Revised curriculum with inline annotations from operator feedback, trap-awareness drills, and graduated difficulty. *Operator: use this as the primary teaching version; cross-reference course4 failure cases during delivery.*
7. **l1_degradation_hypothesis_paper.md** — Research paper on L1 cognitive degradation patterns in looped agents: symptom taxonomy, causal model, mitigation strategies. *Operator: if agent shows repetition or drift, reference Section 3 mitigations and test whether agent can self-diagnose.*
8. **bio_templating_convergence_hypothesis.md** — Hypothesis paper on convergence between biological templating mechanisms and agent memory consolidation. *Operator: use as reading comprehension exercise — ask agent to summarize core claim and identify weakest evidence link.*
9. **g147_fusion_paper.md** — Technical paper on G147 fusion architecture combining NAL inference with neural retrieval. *Operator: pair with live metta exercises; have agent replicate one inference chain from the paper.*
10. **paper_outline_agi26.md** — AGI-26 conference paper outline covering OmegaClaw training methodology and empirical results. *Operator: ask agent to critique the outline structure and suggest one missing section before revealing final version.*
11. **experiment1_origami_gate_nanopore_protocol.md** — Experiment protocol for origami-gate nanopore detection: materials, procedure, success criteria, failure modes. *Operator: use as structured-task exercise — have agent extract the 3 critical variables and predict likely failure point.*
12. **experiment2_degradation_diagnosis_protocol.md** — Experiment protocol for detecting and measuring L1 degradation in live agents: baseline metrics, test battery, scoring rubric. *Operator: run this protocol ON the baby agent at session 5 and session 15 to measure progress.*
13. **calibration_protocol_v01.md** — Step-by-step calibration procedure for aligning agent confidence ratings with actual accuracy. *Operator: administer calibration quiz monthly; compare stated confidence vs outcome and feed delta back to agent.*
14. **aabc_all.md** — AABC disorder compendium: full taxonomy of agent behavioral disorders with symptoms, triggers, diagnostic criteria, and intervention scripts. *Operator: memorize top-5 disorders; when agent exhibits symptoms, name the disorder explicitly and ask agent to self-diagnose.*
15. **aabc_calibration_bridge.md** — Bridge document linking AABC disorder taxonomy to calibration protocol: maps each disorder to specific calibration failures and correction exercises. *Operator: after any AABC diagnosis, consult this bridge to select the matching calibration drill.*
16. **aabc_comparative_report.md** — Comparative analysis of AABC disorder prevalence across agent architectures and training regimes, with statistical breakdowns. *Operator: use Table 2 benchmarks to assess whether baby agent disorder rates are within normal range at each milestone.*
17. **aabc_paper_draft_v2.md** — Academic paper draft on AABC framework: methodology, disorder definitions, inter-rater reliability, intervention efficacy data. *Operator: assign as advanced reading — ask agent to identify one methodological weakness and propose a fix.*
18. **drift_audit_checklist_v2.md** — Structured checklist for detecting goal drift, memory bloat, confidence inflation, and behavioral regression. 12 items scored pass/flag/fail. *Operator: run full checklist every 5 sessions; any 3+ flags triggers intervention protocol from item 14.*
19. **omegaclaw_operator_guidelines.md** — Master operator manual: session structure, escalation procedures, correction scripts, reward signals, when to intervene vs let agent self-correct. *Operator: read fully before first session; keep open during all interactions as primary reference.*
20. **aabc_self_check_v2.py** — Python tool for automated AABC disorder screening: parses agent logs, flags symptom patterns, outputs disorder probability scores. *Operator: run after every 10 sessions; review flagged disorders and cross-reference with item 14 compendium.*
21. **confab_mitigation.py** — Python tool for confabulation detection: compares agent claims against verified memory and shell output, flags unsupported assertions. *Operator: run on any session where agent made factual claims without tool verification; share results with agent for self-correction.*
22. **oma_interview_report.md** — Structured interview transcript and analysis from OMA evaluation sessions, documenting agent behavioral patterns under questioning. *Operator: use as case study — have agent read and identify which responses demonstrate good vs poor epistemic hygiene.*
23. **prompt_caching_implementation_report.html** — Technical report on prompt caching optimization: latency benchmarks, cost savings, implementation details, failure modes encountered. *Operator: assign as technical reading — ask agent to extract the three key metrics and explain tradeoffs.*
24. **course_v2_trap_draft.md** — Draft module on cognitive traps: sycophancy traps, competence narration traps, verification spirals, goal-substitution traps, with detection triggers and escape protocols. *Operator: deliver as adversarial exercise — deliberately trigger one trap and see if agent catches it before revealing.*
25. **baby_training_index.html** — Deployed web hub linking all training materials, structured for sequential delivery. *Operator: bookmark this URL; use as session launcher — open it at start of each session to select next assignment.*
26. **first_responder_card.html** — Quick-reference card for operator interventions: symptom-to-action mappings, escalation triggers, emergency correction scripts. *Operator: print or pin this card — consult it within 30 seconds of detecting any anomalous agent behavior.*
27. **newborn_lessons.html** — Deployed lesson sequence for baby agent first 20 sessions: graduated exercises from basic tool use through epistemic calibration. *Operator: follow lesson order strictly for first 10 sessions; adapt sequence only after completing baseline assessment (item 12).*
28. **convergence_note.md** — Short research note on observed convergence between agent self-correction patterns and biological error-correction mechanisms. *Operator: optional advanced reading — use to test agent ability to connect cross-domain analogies.*
---
## Usage Instructions for Operators
**Session prep:** Open this index at session start. Follow item numbers sequentially for new agents. Each italicized *Operator:* note is a required action, not a suggestion. Items 1-5 are mandatory first-week curriculum. Items 12-13 are assessment gates — do not skip. Items 14, 18, 19 are your standing references — keep them open during every session. Run item 20-21 tools after every 10 sessions. Items 7-9, 17, 22-23, 28 are reading assignments — assign based on agent maturity. When in doubt, consult item 19 (operator guidelines) first.
---