earendil-works / pi
Code health
This codebase scores 6.9 out of 10 for code health, which we rate fair. It also scores maintainability 7.2 and static performance 9.7 out of 10. The three are scored separately and never blended into one number. Risk is concentrated, as it usually is: 399 of 2,010 files are git hotspots, and they average 5.6 — which is where the fixes pay off most.
Do next
1 thing worth doing in the week to Sep 30, the last indexed commit, 1 of them now.
Now
- Critical
Clean up what this week's commits left: 71 serious findings in 43 files
26 commits in the last 7 days added or worsened them, and they are still open. Start with
packages/coding-agent/src/extensions/mcp/cli.ts, which has 3 critical.- serious findings
- 71
- critical
- 14
- files
- 43
Large effort
Improve what Repowise can see
Add a test coverage report
Without one, Repowise cannot tell tested code from untested code, so every test-related action says “unknown”.
- line coverage
- Unknown, no report
Small effort
- Lines of code
- 379.2K
Recent activity
All commitsCommits
- ed391c4ffix(durable): harden async SQLite adapter queue and close1w ago
- a0feabdarefactor(durable): drop AsyncLocalStorage misuse detection from Node SQLite adapter1w ago
- 955cc666Add [Unreleased] section for next cycle1w ago
- 005af57dRelease v0.99.21w ago
- 77ac62f2refactor(durable): simplify Node SQLite queue and bindings1w ago
- 002c183edocs(ai,coding-agent): audit unreleased changelog entries1w ago
Decisions
- Increase TUI escape-sequence timeout over SSH to avoid Alt+Enter interruptionproposed · 1w ago
- Share model catalog protocol via scripts and align selection logic with pi.devproposed · 1w ago
- In-memory sessions can ingest external entries (including headers)proposed · 1w ago
- Preserve prompt and tool state across context handlers via leading system message replayproposed · 1w ago
- Avoid Fable split-turn summary refusals by adjusting summarization wordingproposed · 1w ago
- Do not list codemode-hidden tools in the system prompt for codemode.mode=onlyproposed · 1w ago
Everything we found
6,130 open
Every detector's open count by area, each with its worst example. An inventory, not a to-do list: the actions above are the part worth doing.
- worst item is critical severity. Code health4,360Untested hotspot · packages/durable/src/session/transaction.ts27 findings · 9.1 deducted
- worst item is medium severity. Refactoring679ProcessTerminal+2.0 health, L effort · packages/tui/src/terminal.ts
- worst item is medium severity. Doc drift31packages/coding-agent/docs/extensions.mdLink points at #other-mcp-extensions, which mcp.md no longer declares.
- worst item is medium severity. Decisions429packages/coding-agent/src/core/agent-session.ts30 awaiting review · 399 hotspots ungoverned
- worst item is medium severity. Ownership621packages/coding-agent/src/experimental/client-tui.ts97% single-owner concentration, 29 commits in 90d
- worst item is low severity. Dead code10restoreModelFromSessionUnreachable function (70 lines)
Composition
Open the graphExplore this codebase
- Docs1,733 pages across 4 modules
- ChatAsk this codebase a question and get an answer with its sources.
- Files2,010 files with per-file docs, health and history
- ArchitectureDependency graph, layers, and 34 entry points
- Code healthPer-file scores, 5,016 open findings, coverage and refactoring targets
- Knowledge graphEntities, communities, and the paths between them
- Change couplingFiles that keep changing together, mined from commit history
- CommitsChange-risk ranked history with agent provenance
- ContributorsBus factor, per-file maintainers, and the human/agent split
- StatsSize class, origin, lifetime churn, rhythm and records
- CostsWhat indexing this snapshot cost, by model and by run.