Skip to content

docs: 0.9.0 changelog + blog posts - #3189

Merged
nicoloboschi merged 11 commits into
mainfrom
docs-changelog-0.9.0
Aug 7, 2026
Merged

nicoloboschi merged 11 commits into
mainfrom
docs-changelog-0.9.0

Conversation

@nicoloboschi

@nicoloboschi nicoloboschi commented Aug 5, 2026 •

Copy link
Copy Markdown
Collaborator

Draft — 0.9.0 changelog + blog. Not for merge until the v0.9.0 release is cut.

Consolidates the 0.9.0 docs into one PR (supersedes #3076, closed).

Contents

  • src/pages/changelog/index.md — the 0.9.0 changelog entry. Generated with the real tooling (generate-changelog 0.9.0) against a local-only v0.9.0 tag at origin/main, which was deleted immediately after and never pushed. 40 entries across Breaking Changes / Features / Improvements / Bug Fixes. At actual release the release flow will regenerate this from the pushed tag — treat this as the reviewable preview.
  • blog/2026-08-05-version-0-9-0.md — the usual "What's new in 0.9.0" release post (feature roundup: Knowledge Pages, redesigned control plane, OpenAI Responses provider, per-op llm_extra_body, embedding token cap, reranker failover, mental-model dry-run/keep_trace, Build on Hindsight, upgrade-recommended note).
  • blog/2026-07-29-hindsight-0-9-0.md — the narrative launch post ("memory that pays for itself" — coding-agents plugin + Knowledge Pages, with the sde-bench numbers). Brought over verbatim from docs(blog): 0.9.0 launch post — memory that pays for itself #3076, which is now closed so everything lives in one PR.

Review notes / TODO before release

  • Two 0.9.0 posts, two dates (2026-07-29 from docs(blog): 0.9.0 launch post — memory that pays for itself #3076, 2026-08-05 mine). Align both to the actual release date, and decide the publish order.
  • The launch post still carries docs(blog): 0.9.0 launch post — memory that pays for itself #3076's pending items: hero image /img/blog/hindsight-0-9-0.png, confirm the agentmemorybenchmark.ai "August 2026" date, author-voice pass, flip draft: true.
  • It names opencode, Claude Code, Codex CLI, Gemini CLI, Cursor CLI; the shipped plugin README lists nine harnesses (adds Kilo, Cline, Copilot CLI, Grok Build, Antigravity, Devin — and no Gemini CLI). Reconcile before publish.
  • The generator filed "Require an explicit coding-agents installation target" under Breaking Changes — it's an integration UX change, not a core break. Recategorize if desired.
  • My release post has {/* TODO */} markers for a hero asset and the cross-link to the launch post.
  • Run ./scripts/generate-docs-skill.sh after the real changelog is cut (skill changelog mirror + openapi.json version sync).

@nicoloboschi nicoloboschi changed the title docs: changelog and blog posts for v0.9.0 docs: 0.9.0 changelog + blog posts Aug 5, 2026
@nicoloboschi
nicoloboschi force-pushed the docs-changelog-0.9.0 branch from 75c5858 to c93dcbb Compare August 7, 2026 16:04
The table reported corrections and cost but not how long a task took, which is
the number a reader feels. Computed from the same runs as the other columns —
outputs/sdebench/{hindsight,vanilla}-{claude,codex,opencode}-{1,2,3} on the
benchmark repo's main, meta.wall_s averaged per task across the three runs of
each arm:

  Claude Code   84.7s -> 75.1s  (-11%)
  opencode     174.2s -> 163.3s (-6%)
  Codex CLI     53.6s -> 48.6s  (-9%)

Verified against the published columns before trusting the source: the same
files reproduce corrections 0.85/0.36, 1.20/0.80, 1.34/0.47 and costs -24%,
-13%, -52% exactly.

The prose claim is deliberately the weaker one — every memory run beat its
agent's vanilla AVERAGE, which holds for all nine. 'Faster than every vanilla
run' does not: opencode's slowest memory run (171s) is slower than its fastest
vanilla run (163s), and Codex ties at 51s.
@nicoloboschi
nicoloboschi force-pushed the docs-changelog-0.9.0 branch from c545067 to 521162f Compare August 7, 2026 16:22
@nicoloboschi
nicoloboschi marked this pull request as ready for review August 7, 2026 16:23
@nicoloboschi
nicoloboschi merged commit eb47374 into main Aug 7, 2026
2 checks passed
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.

1 participant