This release adds 3 notable features for engineering teams evaluating rollout.
✓ No known CVEs patched in this version
ReleasePort's take
Light signalTelemetry now persists across Docker volume removal by mirroring counters to a JSON file and can be reset per project via the new UI button.
Why it matters: Ensures telemetry data survives volume cleanup; use Reset Telemetry in each project tile to clear counters while retaining event history.
Summary
AI summaryUpdates Other changes, Highlights, and 60s/project across a mixed release.
Changes in this release
| Type | Severity | Summary | CVE |
|---|---|---|---|
| Feature | Medium |
Telemetry now survives `docker volume rm flaiwheel-data` by mirroring per-project counters to `<docs_path>/.flaiwheel/telemetry.json` and reconstructing state on cold start. Telemetry now survives `docker volume rm flaiwheel-data` by mirroring per-project counters to `<docs_path>/.flaiwheel/telemetry.json` and reconstructing state on cold start. Source: llm_adapter@2026-05-23 Confidence: high |
— |
| Feature | Medium |
Reset Telemetry button added to each per-project tile, zeroing counters via `POST /api/telemetry/reset?project=<name>` while preserving event history. Reset Telemetry button added to each per-project tile, zeroing counters via `POST /api/telemetry/reset?project=<name>` while preserving event history. Source: llm_adapter@2026-05-23 Confidence: high |
— |
| Feature | Medium |
Client Configuration's "VS Code" tab renamed to "VS Code + Copilot" with added GitHub Copilot agent-mode help text. Client Configuration's "VS Code" tab renamed to "VS Code + Copilot" with added GitHub Copilot agent-mode help text. Source: llm_adapter@2026-05-23 Confidence: high |
— |
| Feature | Medium |
Agent instructions updated with "Structured Relations Workflow" section detailing when to add `fixes`, `replaces`, and `depends_on` relations. Agent instructions updated with "Structured Relations Workflow" section detailing when to add `fixes`, `replaces`, and `depends_on` relations. Source: llm_adapter@2026-05-23 Confidence: low |
— |
| Feature | Medium |
Added 9 new tests in `tests/test_telemetry.py` covering mirror writes, rate limiting, cold-start hydration, hot-tier authority, reset semantics, and edge cases. Added 9 new tests in `tests/test_telemetry.py` covering mirror writes, rate limiting, cold-start hydration, hot-tier authority, reset semantics, and edge cases. Source: llm_adapter@2026-05-23 Confidence: low |
— |
| Feature | Medium |
AGENTS.md and install.sh templates now include a Structured Relations Workflow section with rules for fixes, replaces, and depends_on relations. AGENTS.md and install.sh templates now include a Structured Relations Workflow section with rules for fixes, replaces, and depends_on relations. Source: granite4.1:30b@2026-05-23-audit Confidence: low |
— |
| Refactor | Medium |
Indexing functions now skip any path component starting with `.flaiwheel` to prevent internal metadata from polluting index, quality report, or bootstrap output. Indexing functions now skip any path component starting with `.flaiwheel` to prevent internal metadata from polluting index, quality report, or bootstrap output. Source: llm_adapter@2026-05-23 Confidence: low |
— |
| Refactor | Low |
Indexing, quality checks, and bootstrap functions skip paths containing .flaiwheel to avoid indexing internal metadata. Indexing, quality checks, and bootstrap functions skip paths containing .flaiwheel to avoid indexing internal metadata. Source: granite4.1:30b@2026-05-23-audit Confidence: low |
— |
Full changelog
Highlights
- Telemetry now survives
docker volume rm flaiwheel-data. Per-project summary counters are mirrored from the Docker volume into each knowledge repo at<docs_path>/.flaiwheel/telemetry.json. On the next cold start,hydrate_from_mirrors()reconstructs the in-memory state from these files so the Tool Telemetry dashboard does not reset to zero. Hot tier wins when both exist; mirror writes are rate-limited (60s/project) to avoid one Git commit per tool call. Events stay in the volume only (too noisy for the knowledge repo). - Reset Telemetry button on every per-project tile in the Web UI. Zeroes summary counters across both storage tiers via the new
POST /api/telemetry/reset?project=<name>endpoint. Event history is preserved so the rolling 30-day impact metrics keep working. - Agent instructions taught the relations workflow.
AGENTS.mdand both install.sh templates now include a "Structured Relations Workflow" section with three concrete rules (when to addfixes, when to addreplaces, when to adddepends_on). Without this, agents would never know the v3.10.x graph-edge mechanics exist. - Client Configuration's "VS Code" tab is now "VS Code + Copilot" with explicit GitHub Copilot agent-mode help text. The
.vscode/mcp.jsonfile Flaiwheel emits already works for Copilot — no separate snippet needed.
Other changes
indexer._iter_docs,quality._check_*,bootstrap._scan_filesnow skip any path component starting with.flaiwheelso the new mirror file (and any future internal metadata) never pollutes the index, quality report, or codebase-bootstrap output.- 9 new tests in
tests/test_telemetry.pycovering mirror writes, rate limiting, cold-start hydration, hot-tier authority, reset semantics, and edge cases.
Compatibility
Fully backwards-compatible. Existing installs continue to work; mirror files materialise organically on the next save and hydrate on the next cold start. If you don't want telemetry committed to your knowledge repo, add .flaiwheel/ to its .gitignore — Flaiwheel will still read/write the local file but push_pending will skip it.
Update
```bash
bash <(curl -sSL https://raw.githubusercontent.com/dl4rce/flaiwheel/main/scripts/install.sh)
```
Tests
292 → 300, all green.
Full changelog: https://github.com/dl4rce/flaiwheel/blob/main/CHANGELOG.md#3110--2026-05-22
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 dl4rce/flaiwheel
Self-hosted memory and governance layer for AI coding agents. 28 MCP tools with structured knowledge capture, hybrid search (semantic + BM25 + cross-encoder reranking), behavioral documentation nudges, cold-start codebase analyzer, and git-native storage. Single Docker container, zero cloud dependencies.
Related context
Beta — feedback welcome: [email protected]