| name | emerging-pattern-monitor |
| version | 2.2.0 |
| description | Monitors analysis-doc lifecycle in all three lanes — EMERGING→PRODUCTION promotion, PRODUCTION→RETIRING/RETIRED retirement (replacement-readiness as robust community/vendor elements mature), and follow-lane upkeep (followed canons' advance triggers + liveness, per ABSORPTION-MAP.md) |
| triggers | ["review emerging patterns","check pattern promotion","evaluate emerging","promote pattern","review retirement candidates","replacement readiness","what can we retire","follow-lane check","absorption sweep"] |
| auto_load | false |
Emerging Pattern Monitor Skill
Purpose: Monitor EMERGING patterns in analysis documents, track promotion criteria, and recommend when patterns are ready for PRODUCTION status.
When to Use:
- Weekly automated checks (recommended)
- Before quarterly audits
- When Anthropic blog post validates approach
- When production case study published
- User says "review emerging patterns" or similar
What This Skill Does:
- Identifies analysis topics marked as EMERGING in analysis/ docs and SOURCES.md
- Evaluates against promotion criteria
- Tracks evidence accumulation over time
- Recommends promotion when criteria met
- Creates promotion proposals
Promotion Criteria
To promote from EMERGING to PRODUCTION, a pattern must achieve all three:
1. Tier A Source Validation
Requirement: Approach validated by authoritative source
Acceptable sources:
- Anthropic engineering blog post
- Anthropic official documentation endorsement
- Peer-reviewed research (published in ACM, USENIX, IEEE)
- OWASP or NIST standard
Verification:
- Check SOURCES.md for Tier A attribution
- Verify URL is from anthropic.com/engineering or code.claude.com/docs
- Confirm publication date (within 2 years for relevance)
2. Production Validation
Requirement: Pattern used successfully in production
Acceptable evidence:
- 2+ independent case studies from different organizations, OR
- 1 Tier A case study (Anthropic, Fortune 500, named enterprise)
Case study quality criteria:
- Named organization or practitioner
- Quantified outcomes (measurements, time savings, etc.)
- Duration: 30+ days production use
- Publicly verifiable (blog post, conference talk, etc.)
Verification:
- Search SOURCES.md for case study citations
- Check evidence tier (B or higher)
- Verify measurements are specific and dated
3. Community Maturity
Requirement: Pattern or framework shows stability
For frameworks/tools:
- Version 1.0+ released
- Active maintenance (commits/releases within 3 months)
- 100+ GitHub stars OR 10+ contributors
- Documentation exists (README with examples)
For methodology patterns (non-code):
- Referenced by 2+ independent sources
- Stable definition (not rapidly changing)
- Clear usage examples in SOURCES.md
Workflow
Phase 1: Identify EMERGING Topics
Search strategy:
- Grep analysis/ docs for
EMERGING status markers
- Review SOURCES.md for "EMERGING PATTERN" markers
- Check analysis docs with confidence scores below production threshold
Expected output:
EMERGING Topics Found:
1. Reinforcement Learning from Memory (RLM)
- File: SOURCES.md:430-454
- Status: EMERGING since 2025-10-15
2. Claude-Flow multi-agent framework
- File: analysis/orchestration-comparison.md (Framework Selection section, merged 2026-07-16)
- Status: EMERGING since 2025-11-20
Phase 2: Evaluate Each Topic
For each EMERGING topic, create evaluation scorecard:
## Topic: [Name]
**Status**: EMERGING
**First Mentioned**: [date]
### Promotion Criteria Evaluation
#### 1. Tier A Source Validation
- Status: [ACHIEVED / IN PROGRESS / NOT MET]
- Source: [name]
- Evidence Tier: [A/B/C]
#### 2. Production Validation
- Status: [ACHIEVED / IN PROGRESS / NOT MET]
- Case studies: [count] of 2 required
#### 3. Community Maturity
- Status: [ACHIEVED / IN PROGRESS / NOT MET]
- Framework version: [version]
### Overall Assessment
**Promotion Ready?**: [YES/NO] ([count] of 3 criteria met)
**Recommendation**: [Promote / Keep as EMERGING / Downgrade]
Phase 3: Take Action
If promotion ready (all 3 criteria met):
- Update analysis document status to PRODUCTION
- Remove EMERGING markers from documentation
- Update cross-references in related analysis docs
- Document promotion in DECISIONS.md
If blockers exist:
- Document blockers clearly
- Set next review date (1-3 months)
- Monitor for triggering events
Retirement Monitoring (PRODUCTION → RETIRING → RETIRED)
The inverse lane. This project is designed to prune itself as the ecosystem matures (see CONTRIBUTING.md § Retiring a doc) — coverage shrinking is success, not decay — so PRODUCTION docs are also monitored for whether a robust community/vendor element has caught up enough to take over a slice.
Phase R1: Scan PRODUCTION docs for replacement candidates
Triggers to check:
- A new Anthropic first-party feature/command that overlaps a doc's slice (changelog, engineering blog, a new
/-command)
- A community tool reaching GA + real adoption that covers a doc's slice
- A periodic obsolescence sweep — "what did the market just ship that we no longer need to carry?"
Phase R2: Apply the robustness bar (retire only when ALL four clear)
- Supported — first-party (Anthropic) or robustly community-maintained (active, 1.0+, real adoption)
- GA, not preview — generally available, not a research preview/beta
- Covers the slice's substance — does the actual work the doc covered, not a superficial overlap
- Citeable — a stable source to cite (vendor doc, changelog, named feature)
All four clear → recommend RETIRING (defer the slice; keep only what the replacement does not cover). Replacement fully carries the load → RETIRED (tombstone pointer).
Phase R3: Take action (per CONTRIBUTING.md § Retiring a doc)
- Set
status: RETIRING/RETIRED + replacement-by: frontmatter
- Add the Replacement-status banner; register the replacement in SOURCES.md with its boundary (what it does / does not do)
- Update AUDIT-CONTEXT.md routing to defer to the replacement
- Log the retirement in the SOURCES.md refresh log + PLAN.md
Completed retirements: session-quality-tools.md → first-party /insights + native claude doctor (RETIRING 2026-06-04; completed + archived 2026-07-10 — the lane's first full cycle). Watch list: the single watch list now lives in ABSORPTION-MAP.md — one row per routable doc with absorber, lane, robustness-bar state, retained delta, and advance trigger (the former in-file items migrated there 2026-07-16: MCP/skills cost-economics ← /usage became the map's retire-toward row; install-health ← /doctor sits in the map's non-doc watch notes, a slice this project never claimed). Do not maintain a second list here.
Phase F: Follow-lane monitoring (quarterly)
The third lane. Docs with a follows: frontmatter field (see CONTRIBUTING.md § Following a Canon) track an external canon that carries the conceptual load for their slice while the doc keeps only its delta. Quarterly, for each follow-lane row in ABSORPTION-MAP.md:
- Advance-trigger check — has the row's named observable event happened (e.g., the canon's practice productized by a Supported tool, or first-party docs absorbing the slice)? If yes and all four robustness bars now clear, propose converting
follows: → replacement-by: + RETIRING via Phase R3.
- Canon liveness check — is the canon still active (posts/commits within ~6 months)? A frozen canon (the HumanLayer 12-factor precedent: no push since 2025-09) is grounds to recommend dropping the pointer, not to keep following.
- Output — either "no change;
Verified dates bumped for the rows actually re-checked" or a new drafts/ABSORPTION-SCAN-YYYY-QQ.md proposing a wave. A wave is NOT mandatory every quarter (Decision 9's negative precedent: don't over-apply reshaping to well-built docs).
This phase is judgment work and belongs in the quarterly cadence; the weekly review only runs mechanical map-consistency greps (weekly-review step 5b).
Related Analysis
Skill Maintenance
Update this skill when:
- Promotion criteria change
- New evidence tier definitions added
- Anthropic guidance on pattern validation changes
Last Updated: 2026-07-16 (added Phase F follow-lane monitoring; watch list migrated to ABSORPTION-MAP.md as the single store)
Skill Version: 2.2.0