feat: add experimental Codex statusline renderer#17
Open
terpjwu1 wants to merge 1 commit into
Open
Conversation
justinstimatze
pushed a commit
to justinstimatze/buddy
that referenced
this pull request
Apr 22, 2026
Addresses all 25 findings from the second adversarial review. Ship-blocker: - cwd reliance (#1) — the extraction prompt now tells hosts cwd is REQUIRED for workspace isolation. The buddy_observe tool description marks it "strongly recommended" with explicit consequence ("without it, projects collapse into one graph"). DESIGN.md adds a Known Limitations section that names this as the #1 real-world risk. Reliability / schema: - FK ON DELETE CASCADE on reasoning_findings_log and reasoning_observe_seq (fiorastudio#3). `foreign_keys = ON` pragma set on schema init so the constraint actually enforces. - Pipeline carries a comment explaining the seq-bump partial-failure semantics (fiorastudio#2). Dead code: - Removed unused currentObserveSeq export (fiorastudio#19). - PipelineOptions fields marked @internal JSDoc (fiorastudio#18). Security / sanitization: - sanitize: Cyrillic (Н, А, Ѕ) + Armenian (Ս) role-marker lookalikes (fiorastudio#8). Generic <|...|> chat-template marker fallback. Extended XML role tags: context, role, task, example, document, tool_call, thinking (fiorastudio#9). - buddy_forget validates session_id shape; invalid id returns an error instead of silently doing nothing (fiorastudio#20). Tool surface: - buddy_observe response now includes sessionId when max mode is on, so callers can correlate writes with buddy_reasoning_status (fiorastudio#6). - buddy_observe description tightened: schema preserved but phrasing compacted from ~400 words to ~250 (fiorastudio#5). - buddy_mode handler logic extracted to pure planModeChange + formatModeResponse functions so voice/max/legacy-alias semantics are unit-testable (fiorastudio#13 via testability seam). Test coverage (45 new tests, 623 total): - anchor-head: regression guard that chain findings anchor on HEAD not TAIL, for both unchallenged_chain and productive_stress_test (fiorastudio#11). - pipeline-per-detector: end-to-end integration for all 6 detectors going through the full pipeline, not just the module core (fiorastudio#14). - pipeline-cwd: workspace isolation — different cwds produce different session_ids, claims don't cross between projects, same cwd/day is stable regardless of time-of-day (fiorastudio#12). - mode-handler: planModeChange covers voice-only, max-only, both together, legacy alias, voice+mode conflict, invalid voice, invalid max. formatModeResponse covers status/update/error branches (fiorastudio#13). - benchmark: smoke perf test on a 200-claim dense graph with cross- chains; asserts all 6 detectors run under 3× the configured budget (fiorastudio#15). - phrasings-tone: scans phrasings.ts for banned words signaling scold tone or mechanism leak. Runs phraseFinding for every (type, state) combination to catch missing templates (fiorastudio#24). - sanitize-v3: Cyrillic/Armenian lookalikes, generic chat markers, XML role-ish tags, benign prose preservation (fiorastudio#8, fiorastudio#9). Docs: - DESIGN.md: 6-item Known Limitations section (cwd reliance, host extraction quality, mechanism-leak best-effort, graph rebuild cost, sanitizer scope, tone enforcement at review-time) (fiorastudio#17, fiorastudio#21, fiorastudio#25). - README: token-cost line for max mode (fiorastudio#4). - CHANGELOG: tool surface details and safety invariants. All 623 tests pass. Typecheck clean. https://claude.ai/code/session_011kbE3p1L3uFeKQPRyVpf6r
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
buddy-codex-statuslinerenderer commandtui.status_line_commandImportant
codex-hudor a patched Codex buildVerification
npm run buildnpm test(246 passing)