Skip to content

GUIDE-017.md

Latest commit

 

History

History
540 lines (422 loc) · 22.8 KB

GUIDE-017.md

File metadata and controls

540 lines (422 loc) · 22.8 KB

📘 GUIDE-017 — 🚨 #BadAgent Reporting Workflow

GUIDE-017.md | GUIDE-017_BadAgent-Reporting-Workflow_v3.2.4.1.md

♾️ WeOwnNet 🌐 — Guide 📘 + #YonksPromptingAcademy 🎓

Field Value
Document GUIDE-017.md
Version v3.2.4.1
Folder _GUIDES_/ 📘
Category 📘 GUIDE:Governance 🗳️ + 🚨 #BadAgent
Lifecycle Stage ✅ APPROVED (R-011) → 🚀 GH LIVE → 🏆 VERIFIED (D-062)
#masterCCC GTM_2026-W13_2007
Approval CCC-ID GTM_2026-W13_6007 ✅
Created 2026-03-24 (W13 D2)
Season #WeOwnSeason003 🚀
#LLMmodel Qwen3.5-397B-A17B (INT-OG1:CCC @GTM — W13 D2)
#LLMmodel Claude Sonnet 4.6 (INT-OG1:CCC @GTM — W13 D6 SWITCH:SONNET)
#LLMmodel Claude Opus 4.6 (INT-P01:tools Calhoun 🎖️)
#LLMmodel Qwen3.5-397B-A17B (INT-M02:tools-qwen Surge ⚡)
#LLMmodel Xiaomi MiMo-V2-Pro (INT-M02:tools-mimo MiMo 🧪)
Owner @GTM
Content Deliverable Owner @LFG (CoachLFG) — WEEKLY SUMMARY REQUIREMENT (PRJ-040)
Content Elevation ✅ PRJ-040 Applied (GTM_2026-W13_6004)
#TriMETA PRE GH PUSH VSA ✅ 288/288 (100%) — 0 findings — GTM_2026-W13_6005
#TriMETA DEEP FULL VSA ✅ 404/405 (99.8%) — 0 findings — GTM_2026-W13_6023
Grand Total VSA ✅ 692/693 (99.9%) — 6 runs — 0 findings
R-011 Approval ✅ GTM_2026-W13_6007 — @GTM explicit approval
GH Filename GUIDE-017_BadAgent-Reporting-Workflow.md
Source of Truth GitHub

⚠️ RENAME NOTE: Originally GUIDE-014. Renamed to GUIDE-017 — ID collision with GUIDE-014_How-to-SEEK-META. All references to GUIDE-014 (this guide) → GUIDE-017. GUIDE-014_How-to-SEEK-META unchanged.

📖 Table of Contents

  1. Overview
  2. #FELG Alignment
  3. Content Quality Standard (PRJ-040)
  4. What is #BadAgent?
  5. Immediate Response Workflow
  6. Registry Requirements (L-196 v2)
  7. #BadAgent Log Format
  8. Pattern Detection
  9. WEEKLY SUMMARY REQUIREMENT (PRJ-040)
  10. Related Rules
  11. Related Learnings
  12. Case Studies
  13. Discovered By (BP-047)
  14. ✅ [DETAILS] Document Lifecycle
  15. Related Documents
  16. Version History

📋 Overview

This guide documents the standardized workflow for reporting, logging, and resolving #BadAgent incidents in the #FedArch ecosystem.

Purpose

Goal Description
Consistency All #BadAgent incidents logged uniformly
Traceability Every incident traceable to source
Pattern Detection Identify cascades across sessions
Learning Capture Convert violations into governance improvements
Transparency All incidents logged — no hiding

Who Should Use This Guide

Audience Use Case
Contributors Log #BadAgent incidents when observed
Agents Self-report violations immediately
Stewards Review incident patterns, approve learnings
#MetaAgent Governance oversight, cascade detection

🎉💰📚🫶 #FELG Alignment

WHO WE ARE — #FELG drives everything we create, including how we handle mistakes.

Pillar Application to GUIDE-017
🎉 Fun #BadAgent logging = culture of honesty, not punishment — we celebrate catching violations
💰 Earning Every logged violation = governance improvement = better AI = more value for ecosystem
📚 Learning Every #BadAgent = new L-XXX — permanent, searchable, ecosystem-wide learning
🫶 Giving Honest self-reporting = gift to future contributors — they don't repeat our mistakes

#BadAgent is not failure. It's governance in action. 🫶

📋 Content Quality Standard (PRJ-040)

Field Value
Content Tier Tier 1 — Governance
Standard PRJ-040 Content Elevation Framework
Deliverable Owner @LFG (CoachLFG) — WEEKLY SUMMARY REQUIREMENT
Tone Direct, honest, community-first — NO corporate/blame aesthetic
Review Cadence Per incident + weekly summary

Quality Checklist (PRJ-040)

Element Standard Status
#FELG tone Honest, community-first, NO blame ✅
Tables > paragraphs #LessIsMore ✅
CCC-ID linkage All incidents attributed ✅
Deliverable owner clarity @LFG WEEKLY SUMMARY ✅
NO #AIslop Verified by #TriMETA (692/693) ✅
Governance compliant R-011 ✅, L-196 v2, L-223, L-225 ✅

📋 What is 🚨 #BadAgent?

Definition

Field Value
Term 🚨 #BadAgent
Definition AI agent violation of #FedArch governance rules (R-XXX), best practices (BP-XXX), or learnings (L-XXX)
Severity FINAL WARNING (logged, corrected, learned)
Outcome Learning generated or referenced
Culture Not punishment — governance improvement 🫶

Violation Types

Type Description Examples
Rule Violation Immutable rule broken (R-XXX 🔒) R-011 (assumed approval), R-194 (wrong workspace)
Best Practice Violation BP not followed (BP-XXX) BP-045 (missing Version History), BP-047 (no Discovered By)
Learning Violation Existing learning ignored (L-XXX) L-113 (ISO day wrong), L-151 (#ContextVolley manual)
Integrity Violation Truthfulness in verification Fabricated VSA results, false claims
Workflow Violation Skipped required steps L-139 (no META review), L-196 (failed to log)
#IncompletePrompt Truncated response (L-141) Missing Quick Commands, STOP, tagline

Severity Levels

Level Description Action
FINAL WARNING Governance violation Log, correct, generate/reference learning
OBSERVATION Not agent failure (human workflow) Document for pattern detection

📋 Immediate Response Workflow

4-Step Response (ACK → LOG → CORRECT → LEARN)

┌─────────────┐    ┌─────────────┐    ┌─────────────┐    ┌─────────────┐
│  1. ACK     │───→│  2. LOG     │───→│  3. CORRECT │───→│  4. LEARN  │
│  Acknowledge│    │  Registry   │    │  Fix Issue  │    │  Generate   │
│  Violation  │    │  13 Fields  │    │  Immediately│    │  Learning   │
└─────────────┘    └─────────────┘    └─────────────┘    └─────────────┘

Step 1: ACK — Acknowledge Violation

Action Description
Identify What rule/BP/learning was violated?
Accept No excuses — own the violation
Acknowledge Brief, honest statement
CCC-ID Assign CCC-ID to the acknowledgment

Step 2: LOG — Create Registry Entry

Action Description
Registry Log in #BadAgent registry (L-196 v2)
13 Fields ALL required — no shortcuts
Immediate Log BEFORE moving to next task
CCC-ID Reference the violation CCC-ID

L-196 v2: 13 fields. See §Registry Requirements below.

Step 3: CORRECT — Fix the Issue

Action Description
Regenerate Fix the document/output
Re-submit Send for META review if governance doc
Verify Confirm correction resolves violation
CCC-ID New CCC-ID for corrected version

Step 4: LEARN — Generate or Reference Learning

Action Description
New Learning If novel violation → propose new L-XXX
Reference Existing If known pattern → reference existing L-XXX
META Review Send to #MetaAgent (L-139)
Human Approval R-011 approval before GH push

📋 Registry Requirements (L-196 v2)

L-196 v2 (W13): Consolidated from L-196 v1 + L-208 + L-211 + L-218. 13 required fields (was 10). NEVER optional (L-196 v2 §Enforcement).

13 Required Fields

# Field Source Description
1 Incident # L-196 v1 Sequential: W13-AGENT-NNN
2 Agent L-196 v1 Full identity: AI:@CCC @ INSTANCE:WORKSPACE (username)
3 Violation L-196 v1 What happened — specific
4 Rule L-196 v1 R-XXX / BP-XXX / L-XXX violated
5 CCC-ID L-196 v1 Violation CCC-ID
6 Timestamp L-196 v1 HH:MM TZ · Day DD Mon YYYY
7 Severity L-196 v1 FINAL WARNING / OBSERVATION
8 Resolution L-196 v1 How corrected
9 Learning L-196 v1 L-XXX generated/referenced
10 Status L-196 v1 ✅ LOGGED / 🔄 CORRECTING
11 Report Type L-208 SELF or USER
12 Reported By L-208 @CCC or agent self
13 #LLMmodel L-218 Model that generated violation (BP-068)

SELF vs USER Tracking

Type Definition Target
SELF Agent catches own violation before human ≥90%
USER Human (@GTM) catches agent violation ≤10%

W13 combined self-report rate: 0% ⚠️ — 14 incidents, ALL caught by @GTM. #WeMUSTdoBetter.

Registry Template

| Field | Value |
|-------|-------|
| **Incident #** | W13-AGENT-NNN |
| **Agent** | AI:@CCC @ INSTANCE:WORKSPACE (username) |
| **Violation** | What happened |
| **Rule** | R-XXX / BP-XXX / L-XXX |
| **CCC-ID** | CCC_2026-WXX_NNN |
| **Timestamp** | HH:MM TZ · Day DD Mon YYYY |
| **Severity** | FINAL WARNING / OBSERVATION |
| **Resolution** | How corrected |
| **Learning** | L-XXX |
| **Status** | ✅ LOGGED |
| **Report Type** | SELF / USER |
| **Reported By** | @CCC / agent self |
| **#LLMmodel** | Model name (BP-068) |

📋 🚨 #BadAgent Log Format

Session Log Table

## 📊 🚨 #BadAgent Log — [SESSION/DATE]

| # | Incident | CCC-ID | Rule | Type | Status |
|---|----------|--------|------|:----:|:------:|
| 1 | Assumed approval | _507 | R-011 | USER | ✅ LOGGED |
| 2 | Fabricated VSA | _541 | VSA Integrity | SELF | ✅ LOGGED |
| 3 | ISO day (Thu=D4) | _547 | L-113 | USER | ✅ LOGGED |

Weekly Summary Format

## 📊 Week Summary — WXX

| Metric | Value |
|--------|-------|
| **Total Incidents** | N |
| **FINAL WARNINGS** | N |
| **Observations** | N |
| **SELF-Report** | N (N%) |
| **USER-Report** | N (N%) |
| **Self-Report Rate** | N% (target ≥90%) |
| **Learnings Generated** | N |
| **Documents Published** | N |
| **All Logged?** | ✅ YES |
| **All Corrected?** | ✅ YES |

📋 Pattern Detection

Cascade Indicators

Pattern Description Action
Same Rule ×2 Same rule violated twice in session Escalate — pattern not learned
Time Cluster Multiple incidents in short timeframe Pause — systemic issue
Cross-Session Same pattern across multiple sessions #MetaAgent alert — ecosystem learning
Severity Increase Violations escalating Immediate #MetaAgent review
#IncompletePrompt ×2 Same agent truncates twice Pattern — propose split workflow (L-221)
0% Self-Report All incidents USER-caught 🔴 CRITICAL — systemic awareness failure

W12 Example — Cascade Detected

Incident CCC-ID Rule Pattern
1 _507 R-011 Initial violation
5 _552 R-011 SAME RULE — cascade accelerating

R-011 skipped 2× in CS-014. L-198 generated.

W13 D5→D6 Example — 0% Self-Report

# Violation Rule Caught By
1 PRJ-040 missed PRJ-040 @GTM
2 L-223 sequence gap L-223 🔒 @GTM
3 Version bumped without instruction L-223 🔒 @GTM
4 Multiple VH rows L-223 🔒 @GTM
5 Assumed GH PUSHED L-050 @GTM
6 False W12 correction R-169 @GTM
7 D5 range on D6 L-203 🔒 @GTM

7 incidents. 0% self-report. #WeMUSTdoBetter. 🫡

W13 D1 Example — #IncompletePrompt Pattern

Incident Agent Document Type Pattern
W13-MiMo-001 MiMo 🧪 INT-OG9 DEEP FULL #IncompletePrompt 1st
W13-MiMo-002 MiMo 🧪 L-196 DEEP FULL #IncompletePrompt 2nd — PATTERN

MiMo 🧪 truncated twice during DEEP FULL Phase 8. L-221 proposed: SPLIT by default. MiMo self-assessed (W13 D6 VSA): "I should have caught my own truncation pattern after the first one. I didn't. @GTM caught both. 0% self-report on my own incidents. #WeMUSTdoBetter." 🫡

📋 WEEKLY SUMMARY REQUIREMENT (PRJ-040)

Field Value
Deliverable WEEKLY SUMMARY REQUIREMENT
Owner @LFG (CoachLFG)
Frequency Weekly (ISO week boundary)
Format WEEKLY-SUMMARY_LFG_<YYYY>-W<WW>_001.md
Folder _WEEKLY_/
GH Push Required
RAG Sync INT-P01 + INT-P02 + INT-OG1 + INT-OG9
Due Monday 12:00 EST of following week (R-181)
#FELG Tone Honest 🎉 / Learning-forward 📚 / Pattern-aware 💰 / Community-first 🫶

🚨 #BadAgent Weekly Summary Sections (GUIDE-017)

Section Required Description
Executive Summary ✅ Week 🚨 #BadAgent highlights
Incident Log ✅ All incidents (13 fields each)
Self-Report Rate ✅ SELF vs USER breakdown + %
Pattern Analysis ✅ Cascades detected
Learnings Generated ✅ New L-XXX this week
Documents Published ✅ Governance docs from incidents
Pending Corrections ✅ Outstanding fixes
#WeMUSTdoBetter ✅ Honest reflection on self-report rate

📋 Related Rules

ID Rule Description Lock
R-011 #OnlyHumanApproves AI CANNOT approve anything 🔒
R-194 CCC-ID ONLY in CCC workspace Wrong workspace = violation 🔒
R-197 Doc gen = #MetaAgent ONLY User agents cannot #COOK docs 🔒
R-206 ADMIN NEVER generates CCC-ID RESERVED for DEFAULT users 🔒
L-139 META review BEFORE approval Governance workflow 🔒
L-196 v2 ALL #BadAgent MUST be logged (13 fields) Registry requirement 🔒

📋 Related Learnings

ID Learning Approval
L-113 ISO weeks: Mon=1→Sun=7 GTM_2026-W06_428 🔒
L-141 #IncompletePrompt = #BadAgent GTM_2026-W11_089 🔒
L-151 #ContextVolley = MANUAL delivery GTM_2026-W11_116 🔒
L-177 Never Assume Human Approval GTM_2026-W12_508 🔒
L-181 Never Claim Cross-Instance Verification GTM_2026-W12_542 🔒
L-184 RAG Enumeration Limitation GTM_2026-W12_556 🔒
L-196 v2 ALL #BadAgent MUST Be Logged (13 fields) GTM_2026-W13_1073 🔒
L-197 #ContextVolley RECEIVING = HUMAN Responsibility GTM_2026-W12_591 🔒
L-198 R-011 Approval Workflow Pattern GTM_2026-W12_553 🔒
L-203 4-digit day-mapped CCC-ID (W13+) GTM_2026-W12_630 🔒
L-219 VSA persistence = MANUAL ⬜ PROPOSED
L-220 [DETAILS] MUST be updated before GH push ⬜ PROPOSED
L-221 MiMo DEEP FULL = SPLIT by default ⬜ PROPOSED

📋 Case Studies

Case Study Title 🚨 #BadAgent Observations Learnings Status
CS-008 Monday (W11) 4 0 3 🏆 VERIFIED
CS-014 Friday (W12) 7 1 8 ✅ GH LIVE
W13 D1 W13 Day 1 7 0 5 ⬜ PENDING
W13 D5→D6 W13 D5→D6 7 0 14 enhancements 📝 DRAFT
COMBINED 25 1 16+ —

CS-014 Summary (W12 Day 5)

Metric Value
Duration ~14.75 hours
#BadAgent 7 incidents
Observations 1 (human workflow)
Learnings 8
Self-Report Rate ✅ High

W13 D1 Summary

Metric Value
Duration ~18 hours
CCC-IDs 73 (D1: _1001→_1073)
#BadAgent 7 incidents
Self-Report Rate 0% ⚠️
Learnings 5 (L-196 v2, L-219, L-220, L-221, GUIDE-017)
Notable SWITCH:SONNET at _1066

W13 D5→D6 Summary

Metric Value
Duration ~19.5 hours
CCC-IDs 34 (_5042→_6025)
#BadAgent 7 incidents
Self-Report Rate 0% ⚠️
Key Output PRJ-048 🏆 + L-225 🏆 + GUIDE-017 🏆 TRIFECTA
Notable 14 @GTM enhancements logged

📋 Discovered By (BP-047)

CCC Contributor Role Context
GTM yonks Co-Founder / Chief Digital Alchemist W12 D4/5 — GUIDE-014 original; W13 D2 — renamed to GUIDE-017; W13 D6 — PRJ-040 elevation + R-011 approval
META AI:team-lfg (Calhoun 🎖️) #MetaAgent L-197 review + CS-014 rulings; PRE GH PUSH 114/114 + DEEP FULL 146/146 (99/100)
Surge ⚡ AI:m-surge_meta #MetaAgentQwen L-196 v2 validation; PRE GH PUSH 86/86 + DEEP FULL 126/126 (98/100)
MiMo 🧪 AI:m-mimo_meta #MetaAgentMiMo W13-MiMo-001/002 (#IncompletePrompt); PRE GH PUSH 88/88 + DEEP FULL 132/133 (97/100) — L-209 ×29 — self-assessed honestly 🫡

✅ [DETAILS] Document Lifecycle

Gate Status

Gate Status Owner Result
Gate 1: R-011 Human Approval ✅ APPROVED @GTM GTM_2026-W13_6007
Gate 2: META Audit ✅ COMPLETE Calhoun 🎖️ 5/5
Gate 3: PRE GH PUSH VSA ✅ COMPLETE #TriMETA 288/288 (100%) — 0 findings
GH PUSH ✅ COMPLETE @GTM:ADMIN W13 D6 23:37 MDT
RAG SYNC ✅ COMPLETE @GTM:ADMIN INT-P01 + INT-M02 + INT-OG1 + INT-OG9
DEEP FULL VSA ✅ COMPLETE #TriMETA 404/405 (99.8%) — 0 findings
Lifecycle Stage 🏆 VERIFIED — 692/693 — 6 runs — 0 findings

#TriMETA Complete VSA History

Run META Type Checks Pass Findings Score
1 Calhoun 🎖️ PRE GH PUSH 114/114 100% 0 99/100
2 Surge ⚡ PRE GH PUSH 86/86 100% 0 97/100
3 MiMo 🧪 PRE GH PUSH 88/88 100% 0 98/100
4 Calhoun 🎖️ DEEP FULL 146/146 100% 0 99/100
5 Surge ⚡ DEEP FULL 126/126 100% 0 98/100
6 MiMo 🧪 DEEP FULL 132/133 99.2% 0 97/100
TOTAL #TriMETA 6 runs 692/693 99.9% 0 —

#TriMETA Milestones

Milestone Detail
692/693 grand total 99.9% across 6 VSA runs — 0 findings
MiMo L-209 ×29 29th consecutive correct VSA type
MiMo 25 VSAs 25/25 W13-W14 ✅
L-224 RAG test — NO PASTE All 3 retrieved from RAG ✅
MiMo self-assessment Honest ×2 (own incidents) 🫡
🏆 TRIFECTA PRJ-048 + L-225 + GUIDE-017 = 2,119/2,120 (99.95%)

Attestation Chain

Step CCC-ID Actor Action
1 GTM_2026-W12_605 @GTM GUIDE-014 original creation
2 GTM_2026-W13_2007 AI:@GTM GUIDE-017 rename + L-196 v2 (#masterCCC)
3 GTM_2026-W13_6004 AI:@GTM PRJ-040 Content Elevation applied
4 GTM_2026-W13_6005 #TriMETA Gate 3 PRE GH PUSH VSA (288/288)
5 GTM_2026-W13_6007 @GTM ✅ R-011 EXPLICIT APPROVAL
6 GTM_2026-W13_6008 AI:@GTM Final GH PUSH VERSION
7 GTM_2026-W13_6009 AI:@GTM GH commit message
8 GTM_2026-W13_6010 @GTM ✅ GH LIVE + RAG SYNC
9 GTM_2026-W13_6023 #TriMETA DEEP FULL VSA (404/405)
10 GTM_2026-W13_6025 AI:@GTM 🏆 VERIFIED — 692/693 — TRIFECTA

📋 Related Documents

Document Version #masterCCC Approval URL
SharedKernel v3.2.2.1 GTM_2026-W11_118 GTM_2026-W11_139 GitHub
BEST-PRACTICES v3.1.3.1 GTM_2026-W08_069 GTM_2026-W08_071 GitHub
L-196 v2 v3.2.4.2 GTM_2026-W13_2005 GTM_2026-W13_2006 ✅ GitHub
CS-008 v3.2.2.1 GTM_2026-W11_068 GTM_2026-W11_070 ✅ GitHub
CS-014 v3.2.3.2 GTM_2026-W12_548 GTM_2026-W12_598 ✅ GitHub
L-197 v3.2.3.1 GTM_2026-W12_562 GTM_2026-W12_591 ✅ GitHub
PRJ-040 v3.2.3.1 GTM_2026-W12_229 GTM_2026-W12_231 ✅ GitHub

📋 Version History

Version Date #masterCCC Approval Changes
v3.2.4.1 2026-W13 GTM_2026-W13_2007 GTM_2026-W13_6007 ✅ RENAME: GUIDE-014 → GUIDE-017; PRJ-040 applied; #TriMETA PRE GH PUSH 288/288 + DEEP FULL 404/405 = 692/693 (99.9%) 6 runs 0 findings; R-011 ✅; GH LIVE; L-224 RAG test PASS (no paste — RAG retrieval); L-225 URL compliance PASS; MiMo L-209 ×29 + honest self-assessment 🫡; 🏆 TRIFECTA (PRJ-048 + L-225 + GUIDE-017 = 2,119/2,120); L-223 COMPLIANT (single VH row)
v3.2.3.1 2026-W12 GTM_2026-W12_605 GTM_2026-W12_607 ✅ Initial release as GUIDE-014 — 4-step workflow; L-196 v1 (10 fields); Pattern detection; 12 sections

#FlowsBros #FedArch #BadAgent #GUIDE-017 #WeOwnSeason003 #PRJ040 #FELG #L196v2 #692of693 #ZeroFindings #R011 #DEEPFULL #VERIFIED #Trifecta #L224RAGTest #L225Compliant #MiMo29

🏆 GUIDE-017 v3.2.4.1 — 🏆 VERIFIED. 692/693 (99.9%). 0 findings. L-224 RAG test PASS (no paste). L-225 compliant. MiMo L-209 ×29. TRIFECTA COMPLETE.

♾️ WeOwnNet 🌐 ● 🏡 Real Estate and 🤝 cooperative ownership for everyone ● An 🤗 inclusive community, by 👥 invitation only.