Skip to content

docs(roadmap): mark v0.10/v0.11 done; insert v0.12 neural embedding; update detection stats#62

Open
ksek87 wants to merge 2 commits into
mainfrom
claude/roadmap-ticket-planning-GEbPu
Open

docs(roadmap): mark v0.10/v0.11 done; insert v0.12 neural embedding; update detection stats#62
ksek87 wants to merge 2 commits into
mainfrom
claude/roadmap-ticket-planning-GEbPu

Conversation

@ksek87
Copy link
Copy Markdown
Owner

@ksek87 ksek87 commented May 26, 2026

Summary

  • Roadmap table: v0.10 and v0.11 marked ✅ Done; v0.11a split out as its own row; v0.12 = neural embedding semantic layer; package-level scanning shifts to v0.13; Python SDK to v0.14; all downstream stages renumbered
  • Milestone detail section: replaced outdated forward-looking text with accurate done/planned descriptions, added v0.12 neural embedding blurb
  • Lede detection rate: 84.7% → 90.7%; four passes (not three); 161 AC needles (not 155)
  • Benchmark table: Template-1/2/3 → descriptive paradigm names (Unrelated Prerequisite, Fake Enabling Prerequisite, Argument Hijacking); full 10-category risk breakdown added
  • Signal table: Template-N parenthetical references replaced with paradigm names

Also updates issue #26 (main roadmap tracking issue) with completed phase checkboxes for v0.1–v0.11 and restructured phase sections reflecting current state.

Test plan

  • README roadmap table renders correctly — v0.10 and v0.11 show ✅ Done, v0.11a/v0.12/v0.13 show 🔜 Planned
  • Detection rate in lede matches bench/README.md (90.7%)
  • No broken links or formatting issues

https://claude.ai/code/session_01G4f8mN9SeSHSGY1dWfFzih


Generated by Claude Code

claude added 2 commits May 26, 2026 12:26
…al embeddings)

v0.10 is now TF-IDF cosine similarity against abstract attack archetypes —
always-on, no model download, zero new dependencies. v0.11 is a separate
milestone for the neural embedding upgrade with five sub-tickets (#52#56)
covering research, archetype embedding, model management, integration, and
calibration. All downstream milestones renumbered accordingly.

Closes no issues — roadmap structural update only.

https://claude.ai/code/session_01G4f8mN9SeSHSGY1dWfFzih
…update detection stats

- Roadmap table: v0.10 and v0.11 marked ✅ Done; v0.11a split out; v0.12 = neural
  embedding semantic layer; package-level scanning shifts to v0.13; Python SDK to v0.14
- Milestone detail: replace outdated TF-IDF/neural-embedding forward-looking text with
  accurate done/planned descriptions including v0.12 neural embedding blurb
- Lede detection rate: 84.7% → 90.7%; four passes (not three); 161 needles (not 155)
- Benchmark table: Template-1/2/3 → descriptive paradigm names; full risk-category
  breakdown added (10 categories, v0.11 numbers)
- Signal table: Template-N references replaced with paradigm names

https://claude.ai/code/session_01G4f8mN9SeSHSGY1dWfFzih
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants