Pull Request Explorer
| Title | Author | Size | AI | Cycle Time | Review | Merged |
|---|---|---|---|---|---|---|
| fix(agent): session.id on every metric point; keep a metrics window whose export failed | locnguyen1986 | L | AI | 147.6h | 2.5h | Oct 05, 2026 |
|
LLM analysis pending This PR has not been analyzed yet. |
||||||
| fix(windows): retarget shortcuts when the main binary is renamed on update | qnixsynapse | M | No | 3.8h | 0.1h | Oct 05, 2026 |
|
LLM analysis pending This PR has not been analyzed yet. |
||||||
| fix(linux): make the AppImage .DirIcon relative to the AppDir | qnixsynapse | S | No | 0.9h | 0.0h | Oct 05, 2026 |
|
LLM analysis pending This PR has not been analyzed yet. |
||||||
| feat(agent): Hide Secrets - keep credentials out of model requests | thinhlpg | XL | No | 95.5h | 1.1h | Oct 05, 2026 |
|
LLM analysis pending This PR has not been analyzed yet. |
||||||
| fix(i18n): translate toolCall strings for all locales and guard tools.json completeness | jamesdam | XL | No | 0.6h | 0.1h | Oct 02, 2026 |
|
LLM analysis pending This PR has not been analyzed yet. |
||||||
| ci: pin Rust toolchain to 1.98.0 via rust-toolchain.toml | jamesdam | S | No | 1.9h | 0.1h | Oct 02, 2026 |
|
LLM analysis pending This PR has not been analyzed yet. |
||||||
| chore(flatpak): build 0.8.5 with the in-process CUDA + Vulkan llama.cpp engine | qnixsynapse | L | No | 19.2h | 19.1h | Oct 02, 2026 |
|
LLM analysis pending This PR has not been analyzed yet. |
||||||
| fix(tooling): convert i18n completeness scripts to ESM | jamesdam | S | No | 0.3h | 0.0h | Oct 02, 2026 |
|
LLM analysis pending This PR has not been analyzed yet. |
||||||
| docs(qa): update pre-release checklist for v0.8.5 and remove stale autoqa copy | jamesdam | L | AI | 2.8h | 1.2h | Oct 02, 2026 |
|
LLM analysis pending This PR has not been analyzed yet. |
||||||
| fix(clippy): remove redundant #[must_use] on write_background_output | jamesdam | XS | No | 2.4h | 0.0h | Oct 02, 2026 |
|
LLM analysis pending This PR has not been analyzed yet. |
||||||
| feat(cli): run-time provider overrides for gateways (base URL, custom headers, User-Agent, X-Session-Id) | locnguyen1986 | XL | AI | 44.7h | 2.7h | Oct 01, 2026 |
SummaryFeatureAdd runtime provider overrides for gateways Enables Jan Agent to point to model gateways at runtime, allowing dynamic configuration without editing config files and improving integration flexibility for deployments such as Tokamak. Health Assessment
X-Large
High
Medium
AI Details
Usage:
Authored by AI
Category:
Code AI
Tools:
Claude
Confidence:
0.95
Tech Stack
Languages:
Rust
|
||||||
| feat(agent): steer and stop subagents, freeze the session start into the system prompt | qnixsynapse | XL | No | 25.9h | 23.0h | Oct 01, 2026 |
SummaryFeatureSteer subagents and freeze session start Adds subagent steering and session‑start snapshot, improving control over subagent lifecycles and ensuring consistent system prompts across a session. Health Assessment
X-Large
High
Medium
AI DetailsTech Stack
Languages:
Rust, Typescript, Javascript, Json
|
||||||
| feat(cli): experimental `jan acp` Agent Client Protocol server | qnixsynapse | XL | No | 11.0h | 9.9h | Oct 01, 2026 |
SummaryFeatureAdd experimental ACP server for Jan CLI Enables external editors like Zed and JetBrains to drive Jan via ACP, expanding integration options. Health Assessment
X-Large
Medium
Medium
AI Details
Confidence:
0.95
Tech Stack
Languages:
Rust
|
||||||
| ci(rust-check): lint the standalone jan-cli crate | qnixsynapse | S | No | 1.1h | 0.3h | Sep 30, 2026 |
SummaryCI/CDAdd CI lint for standalone jan-cli crate Ensures the jan-cli Rust crate is linted in CI, catching warnings early and improving code quality. Health Assessment
Small
Low
Low
AI Details
Confidence:
0.20
Tech Stack
Languages:
Rust
|
||||||
| feat(ci): ship windows arm64 on the stable channel | Minh141120 | M | No | 0.9h | 0.4h | Sep 30, 2026 |
SummaryCI/CDShip windows arm64 on stable channel Adds support for building and publishing Windows ARM64 binaries in the CI pipeline, ensuring stable releases include the new architecture without gating other platforms. Health Assessment
Small
Low
Low
AI DetailsTech Stack
Languages:
Yaml
Frameworks:
Github-Actions
|
||||||
| fix(docs): bring back docs search and make the /docs search field work | thinhlpg | M | No | 2.0h | 1.9h | Sep 29, 2026 |
SummaryBug FixRestore docs search and fix /docs field Restores search functionality on the docs site, improving user navigation and content discoverability. Health Assessment
Small
Low
Low
AI Details
Confidence:
0.10
Tech Stack
Languages:
Typescript
Frameworks:
React
|
||||||
| fix(mcp,tools): do not pass gateway credentials to child processes | locnguyen1986 | L | AI | 0.2h | - | Sep 29, 2026 |
SummaryBug FixPrevent credential leakage to child processes Ensures sensitive gateway credentials are not exposed to spawned processes, improving security for MCP sessions. Health Assessment
Large
Low
Low
AI Details
Usage:
Authored by AI
Category:
Code AI
Tools:
Claude
Confidence:
0.95
Tech Stack
Languages:
Rust
Frameworks:
Tauri
|
||||||
| feat(agent): OTel delta temporality, export retry, bounded series, and signal flush | thinhlpg | XL | No | 1.8h | 1.8h | Sep 29, 2026 |
SummaryFeatureAdd OTel telemetry improvements and signal flush Enhances telemetry reliability and efficiency by adding delta temporality, export retry, bounded series, and graceful shutdown on signals, improving observability and reducing data overhead. Health Assessment
X-Large
Low
Low
AI Details
Confidence:
0.20
Tech Stack
Languages:
Rust
|
||||||
| feat(llamacpp): upgrade the engine to 0.5.0 (b11146) | qnixsynapse | M | No | 1.8h | 1.2h | Sep 29, 2026 |
SummaryFeatureUpgrade llama.cpp engine to 0.5.0 Upgrades the in‑process llama.cpp engine to the latest release, improving performance, compatibility, and enabling new features while keeping build consistency. Health Assessment
Small
Low
Low
AI Details
Confidence:
0.95
Tech Stack
Languages:
Rust, Typescript, Cpp
|
||||||
| feat(agent): opt-in OpenTelemetry (OTLP) metrics + logs export | thinhlpg | XL | No | 23.5h | 23.5h | Sep 29, 2026 |
SummaryFeatureOpt‑in OTLP metrics and logs export Adds optional OpenTelemetry export for agent metrics and logs, enabling richer observability and integration with Tokamak Insights while preserving privacy and performance. Health Assessment
Large
Medium
Low
AI Details
Confidence:
0.80
Tech Stack
Languages:
Rust, Python, Shell
|
||||||
| feat(agent): AGENTS.md fallback, slash suggestions mid-run, and /vibe-setting | thinhlpg | XL | No | 23.7h | 4.7h | Sep 29, 2026 |
SummaryFeatureAdd AGENTS.md fallback, slash suggestions, and vibe setting Users can now use AGENTS.md as a default configuration, receive slash command suggestions while a run is live, and adjust settings via a new /vibe-setting command, improving workflow efficiency. Health Assessment
X-Large
Medium
Low
AI Details
Confidence:
0.80
Tech Stack
Languages:
Rust
|
||||||
| test(rpc): stream-json parity and session isolation (tests only) | thinhlpg | XL | No | 24.0h | 5.9h | Sep 29, 2026 |
SummaryTestAdd RPC parity and session isolation tests Adds tests verifying RPC and stream-json parity and ensuring session isolation, improving reliability of RPC interactions. Health Assessment
X-Large
Low
Low
AI DetailsTech Stack
Languages:
Rust
|
||||||
| fix(agent): show upstream retries live in the TUI and headless output | qnixsynapse | L | No | 13.9h | 13.4h | Sep 29, 2026 |
SummaryHealth AssessmentAI Details
Confidence:
0.75
Tech Stack
Languages:
Python, Typescript, Javascript, Rust
|
||||||
| fix(agent-tui): keep a batched edit's diff under its own call row | qnixsynapse | M | No | 13.0h | 13.0h | Sep 29, 2026 |
SummaryBug FixFix batched edit diff placement in TUI Ensures edit diffs display next to their call rows, improving user clarity in the TUI. Health Assessment
Medium
Low
Low
AI DetailsTech Stack
Languages:
Rust
|
||||||
| fix(agent): stop early compaction on 1M models and report every compaction | qnixsynapse | XL | AI | 2.9h | 1.1h | Sep 28, 2026 |
SummaryBug FixFix early compaction on 1M models Improves session budget handling and compaction reporting, preventing premature history truncation for large-context models. Health Assessment
X-Large
High
Low
AI Details
Usage:
AI Assisted
Category:
Code AI
Tools:
Claude, ChatGPT
Confidence:
0.95
Tech Stack
Languages:
Rust, Typescript, Javascript, Python
|
||||||
| feat(agent): user-scope skills in ~/.jan/skills, readable plugin skills, and symlinked skill folders | thinhlpg | XL | No | 4.2h | 4.1h | Sep 28, 2026 |
SummaryFeatureAdd user-scope skills and plugin skill visibility Enables user‑scoped skills accessible across projects, improves skill visibility, and supports symlinked skill folders, enhancing flexibility for CLI and TUI. Health Assessment
Medium
Low
Low
AI Details
Confidence:
0.75
Tech Stack
Languages:
Rust
|
||||||
| fix(agent): follow-ups to host delegation and systemPrompt (#9081) | qnixsynapse | L | No | 0.4h | 0.2h | Sep 28, 2026 |
SummaryBug FixFix host delegation and systemPrompt behavior This PR corrects memory handling, system prompt persistence, and subagent name resolution for host‑prompted sessions, ensuring reliable agent behavior and consistent thread metadata. Health Assessment
Large
Medium
Low
AI Details
Confidence:
0.80
Tech Stack
Languages:
Python, Typescript, Rust
|
||||||
| fix(agent): exact wire names win over declared host names; correct the RPC docs | qnixsynapse | L | No | 1.1h | 0.3h | Sep 28, 2026 |
SummaryBug FixFix agent host name collision and update RPC docs Corrects host name resolution logic to prioritize exact wire names, resolves documentation inconsistencies, and adds comprehensive tests to ensure reliable agent behavior. Health Assessment
Medium
Low
Low
AI Details
Confidence:
0.95
Tech Stack
Languages:
Python, Rust, Typescript
|
||||||
| feat(agent): host-tool-only delegation and a verbatim host system prompt over RPC | qnixsynapse | XL | No | 0.6h | 0.4h | Sep 28, 2026 |
SummaryFeatureAdd host-tool-only delegation and system prompt RPC This update introduces host-tool-only delegation, restricting subagent tool access to the host's allowed set. It also adds a system prompt field that can be sent verbatim over RPC, giving callers full control over the agent's initial context. Health Assessment
X-Large
Medium
Low
AI Details
Confidence:
0.75
Tech Stack
Languages:
Python, Typescript, Javascript, Rust
|
||||||
| feat(agent): /reload re-reads agent.toml, skills, plugins and JAN.md mid-session | thinhlpg | XL | No | 1.6h | 0.3h | Sep 28, 2026 |
SummaryFeatureAdd /reload command to refresh agent config mid-session Enables users to reload configuration, plugins, skills, and system prompt on the fly, eliminating the need to restart the application and improving productivity. Health Assessment
X-Large
Low
Low
AI DetailsTech Stack
Languages:
Rust
|
||||||