research.ietf — Status

2026-05-22 — Deployed at ietf.nennemann.de + full data refresh

Deployed the Flask dashboard to nennemann-dev behind Caddy: https://ietf.nennemann.de (basic_auth vorschau, noindex, production mode). Docker dev-ietf-1 on :8082→:5000, data dir mounted read-write. Dockerfile fixed to bind 0.0.0.0. Infra in nennemann-biz/infra/dev (compose + Caddy vhost).
Data refresh (first since 2026-03-08): 761→889 drafts. Full Sonnet pipeline (rate/embed/ideas/score/gaps/convergence). Live stats: 816 relevant, 722 authors, 973 ideas, 18 gaps, 170 convergent. Fresh drafts.db synced to server.
New tool: scripts/backfill-unrated-text.py (fetch leaves new drafts without full text — rev-suffix issue).
Bug fix: db.insert_gaps() now FK-safe (was a blanket DELETE FROM gaps).

Refresh procedure documented in workspace docs/services.md (sync only drafts.db; analytics.db is server-owned).
ANTHROPIC_API_KEY shell env was stale — use the key in .env.
Optional: schedule recurring refresh (scripts/scheduled-update.sh) if desired.
16 drafts yielded no extractable ideas (short/non-technical) — expected, left as-is.

Go ValidateHashFormat updated to -01 spec: Was still validating old algorithm:base64url format (e.g. sha-256:abc...). Now validates plain base64url without prefix, matching -01 spec and RFC 9449.
Go ErrHashFormat error message updated to match new format.
IMPROVEMENTS.md updated: Section 5 and section 6 hash format entries corrected.
Draft rebuilt: build.sh run with kramdown-rfc2629 (freshly installed) and xml2rfc. XML/TXT/HTML regenerated.
All tests pass: Python (56 tests, 90% coverage), Go (all pass).

All claim name changes were already implemented in core library code (types, create, verify, dag) for both Go and Python:

Change	Status
`par` -> `pred`	Done (both)
`pol`/`pol_decision` removed from core, moved to `ect_ext`	Done (both)
`sub` removed	Done (both)
`typ: wimse-exec+jwt` -> `exec+jwt` (preferred)	Done (both)
`MaxParLength` -> `MaxPredLength`	Done (both)
Hash format: `algorithm:base64url` -> plain base64url	Fixed in this session (Go was still using old format)

Primary draft: draft-nennemann-wimse-ect-01 (Execution Context Tokens for Distributed Agentic Workflows)

The -01 draft is structurally complete and IETF-compliant:

All required sections present: Abstract, Introduction, Conventions/Definitions, Security Considerations, Privacy Considerations, IANA Considerations, References (normative + informative)
BCP 14 (RFC 2119/8174) boilerplate included
IANA registrations complete: 2 media types, 1 HTTP header field, 6 JWT claims
3 use cases in appendix (L1, L2, L3)
Comprehensive Related Work section (WIMSE, OAuth, TxnTokens, OpenTelemetry, PROV, SCITT, RATS, emerging agent protocols)
No TODOs/FIXMEs/TBDs in the draft

Blog post outdated claim names (FIXED): par -> pred, ext -> ect_ext in blog-ect-assurance-levels.md
Refimpl README outdated (FIXED): Added migration note documenting -00 vs -01 claim name changes, updated specification references
Refimpl IMPROVEMENTS.md (FIXED): Added section 6 documenting all -01 migration tasks needed

Refimpl not updated to -01: Reference implementations (Go + Python) still use -00 claim names (par instead of pred, pol/pol_decision instead of ect_ext, sub not in spec). This is the biggest gap before submission.
Build tooling: kramdown-rfc2629 not installed in this environment. Need Ruby gem to rebuild from markdown.
Date stale: XML was built 2026-03-06. Needs rebuild before submission.
Other draft workspace: workspace/drafts/ contains 7 other draft directions (gap-analysis, agent-context-policy, new-drafts, etc.) — none are as mature as the ECT draft.