This release adds 2 notable features for engineering teams evaluating rollout.
✓ No known CVEs patched in this version
Topics
+6 more
Summary
AI summaryAdded longhand demo CLI command with sandboxed sample corpus.
Changes in this release
| Type | Severity | Summary | CVE |
|---|---|---|---|
| Feature | Low |
Adds `longhand demo` CLI command for sandboxed preview. Adds `longhand demo` CLI command for sandboxed preview. Source: llm_adapter@2026-07-15 Confidence: high |
— |
| Feature | Low |
Adds `longhand.demo` package with corpus generator and runner. Adds `longhand.demo` package with corpus generator and runner. Source: llm_adapter@2026-07-15 Confidence: high |
— |
| Feature | Low |
Adds 6 new tests verifying deterministic corpus and recall functionality. Adds 6 new tests verifying deterministic corpus and recall functionality. Source: llm_adapter@2026-07-15 Confidence: high |
— |
Full changelog
longhand demo — try Longhand on a sample corpus without touching your real history.
Added
- New
longhand demoCLI command. Generates a sandboxed store at
/tmp/longhand-demo-<timestamp>/, seeds it with 3 fictional Claude Code
sessions covering a realistic Stripe-webhook bug + Supabase auth migration- downstream 401 fix on a
demo-shopproject, then walks throughrecall
andrecall_project_statusagainst the seeded store so the user can see
what the output looks like before pointing Longhand at their own data.
Cleans up afterwards; pass--keepto leave the sandbox in place for
further exploration withLONGHAND_DIR=<path> longhand ….
- downstream 401 fix on a
longhand.demopackage containing the corpus generator
(longhand/demo/corpus.py) and walkthrough runner
(longhand/demo/runner.py). 6 new tests verify the corpus is
deterministic, valid Claude Code event shape, ingests cleanly into a
fresh store, and that recall returns sensible results against the seeded
data.
Why
The 2026-05-17 audit identified an onboarding gap: new users had no way to
preview Longhand on safe data before running longhand setup on their
real ~/.claude. longhand demo closes that gap with a 60-second
walkthrough that exercises the cross-session recall, file history, and
project-status surfaces without modifying anything on disk outside the
sandbox.
Tests
- 228 tests passing (was 222).
Weekly OSS security release digest.
The CVE patches and breaking changes that affected production tools this week. One email, every Sunday.
No spam, unsubscribe anytime.
Share this release
About Wynelson94/longhand
Persistent local memory for Claude Code. Indexes every session JSONL verbatim into SQLite + ChromaDB for semantic recall (~126ms)
Related context
Related tools
Earlier breaking changes
- v0.13.0 Deprecates several CLI commands; they will be removed at v1.0.
Beta — feedback welcome: [email protected]