Skip to content

research(aif-handoff-bridge): Sub-wave B — Variant B filesystem-watcher - #267

Merged
artyhoo merged 1 commit into
stagingfrom
research/aif-handoff-bridge-variant-b
May 29, 2026
Merged

research(aif-handoff-bridge): Sub-wave B — Variant B filesystem-watcher#267
artyhoo merged 1 commit into
stagingfrom
research/aif-handoff-bridge-variant-b

Conversation

@artyhoo

@artyhoo artyhoo commented May 28, 2026

Copy link
Copy Markdown
Owner

Summary

Sub-wave B R-phase of the aif-handoff-as-runtime-bridge umbrella — evaluates whether aif-handoff supports (or could feasibly gain) a filesystem-watcher/directory-watch mode for auto-ingesting .claude/orchestrator-prompts/<umbrella>/kickoff.md files written by /meta-orchestrator.

Verdict: REJECT — Variant B is not actionable now.

Key findings:

  • No apps/coordinator/src/watchers/ directory exists in lee-to/aif-handoff (confirmed via 4 DeepWiki probes + full file-tree search)
  • The fs/watch protocol returned by DeepWiki probe 1 is a Codex App Server client-side notification protocol (packages/runtime/src/adapters/codex/appServer/generated/) — NOT aif-handoff's coordinator task-ingestion
  • packages/agent/src/taskWatchdog.ts is a stale-task recovery watchdog, not a directory watcher
  • Zero PRs/issues proposing inotify/fsnotify/chokidar/directory-watch task ingestion in lee-to/aif-handoff
  • Implementing Variant B would require ~400-800 LOC upstream PR to aif-handoff
  • DECISION=C zero-infra hard constraint violated by Docker+daemon requirement

Admission gate §2.4 re-sweep: packages/mcp/package.json changed (SDK version pin ^1.12.11.29.0) — dependency bump only, no new MCP tools shipped.

Evidence channels (T1 sampling floor ≥5 — 9 total)

  1. DeepWiki probe 1: filesystem watcher existence
  2. DeepWiki probe 2: task creation paths
  3. DeepWiki probe 3: daemon/process-management
  4. DeepWiki probe 4: wiki structure (no watcher//daemon//ingest/ dir)
  5. DeepWiki probe 5 (adversarial counter-probe, T7): explicit no-watcher confirmation
  6. gh api git/trees/main?recursive=1 file-tree search
  7. gh search prs for inotify/fsnotify/chokidar/watcher (0 results for dir-watch)
  8. gh api issues sweep (50 issues, none mention dir-watch)
  9. WebSearch ×3 phrasings on «filesystem watcher agent orchestration» (no production pattern found)

Scope note

Markdown-only artefact: docs/meta-factory/research-patches/2026-05-29-aif-handoff-bridge-variant-b-watcher.md (272 lines). No code edits, no skill edits, no install.sh changes.


§1.7 Forward-check applied

§1.7 Backward-check applied

Falsifier

If lee-to/aif-handoff ships packages/agent/src/dirWatcher.ts (or equivalent) that polls a configurable filesystem path for new task-definition .md files and creates tasks via createTask — re-evaluate Variant B as ADAPT.

…er viability probe

Verdict: REJECT — no directory-watch task-ingestion capability exists in
aif-handoff (confirmed via 4 DeepWiki probes + full file-tree search +
gh issue/PR search); upstream PR effort ~400-800 LOC; DECISION=C zero-infra
hard constraint violated by daemon requirement. §8 stop condition fires.

Prior-art: skipped — research-patch markdown only, no capability commit per CLAUDE.md (no new dep, no new file under packages/, LOC count does not apply to docs/)
@artyhoo
artyhoo merged commit a49244f into staging May 29, 2026
22 checks passed
artyhoo added a commit that referenced this pull request May 29, 2026
#269)

Mechanical corrections to merged SW-A (#268) + SW-B (#267) research-patches per
Phase -1 Stage 1→2 cold-review. No verdict changes — SW-A REFERENCE and SW-B
REJECT both unaffected.

- B1: SW-B :39, :117 — PR #128 IS merged. Original record used jq camelCase
  '.mergedAt' which returns null; the actual snake_case '.merged_at' returns
  2026-05-26T08:07:12Z (merge SHA 51ce96a). Verified 2026-05-29 via
  'gh api repos/lee-to/aif-handoff/pulls/128 --jq .merged_at'.
- M1: SW-A :343 — gate-4 evidence said "all are from PR #128" but BOTH
  PR #127 (e1ffa70, merged 2026-05-15) AND PR #128 (51ce96a, merged
  2026-05-26) are in the 30-day window. Both touch packages/runtime/ only.
- m1: SW-A :392 — Probe 6 URL hash typo ae7472a9 -> ae7072a9 (aligned to
  line 119 body-text canonical version; inferred from internal consistency,
  not externally re-verified due to DeepWiki client-side rendering).
- m2: SW-B :210 — overclaim "≥4 distinct phrasings on dir-watch" corrected
  to actual breakdown: 2 direct + 2 adjacent DeepWiki + 1 file-tree +
  1 gh + 3 WebSearch = 9 channels. Multi-channel triangulation noted.
- m3: SW-B §0 falsifier (:17) — aligned to §5 specificity ("…and creates
  tasks via the internal createTask function — re-evaluate as ADAPT with
  three criteria 1/2/4 re-scored…").

Phase -1 Stage 1→2 cold-review verdict was REVISE; this patch closes
B1+M1+m1/m2/m3 dispositions per state.md §3.4. DN-1 and DN-2 remain
maintainer-call; SW-C dispatch still blocked on those answers.

Prior-art: skipped — mechanical text correction, no capability commit, no
new dependency, no new file ≥50 LOC. Per CLAUDE.md capability-commit
definition, this is a refactor/typo-fix grade change to existing artifacts.
artyhoo added a commit that referenced this pull request May 29, 2026
… Implementer-equivalent only) value-add audit (#276)

R-phase patch for Sub-wave C of the aif-handoff-as-runtime-bridge umbrella.
Evaluates Variant C (kickoff §3 lines 124-145): aif-handoff as Implementer-
equivalent only, bypass Planner+Reviewer cycle, thin CLI wrapper for kickoff
dispatch + kanban status tracking.

Verdict: REJECT (BFR-default §1 ladder). Rationale:
- The kickoff-framed "aif-handoff exec --kickoff <path>" CLI does not exist
  in lee-to/aif-handoff (DeepWiki probes 1+5, 2026-05-29).
- No first-class Implementer-only mode; skipReview:true bypasses Reviewer but
  Planner is mandatory unless accept_existing_plan with on-disk PLAN.md
  (same disk coupling SW-A flagged for Variant A).
- BEFORE/AFTER maintainer-action count: 25% literal / 0% cognitive reduction
  (T-AIF-BRIDGE-C table §4) — below kickoff §8 STOP 30% threshold → verdict
  "Variant C value-add insufficient".
- Pure-tracker pattern (paused:true + autoMode:false + manual state-machine
  transitions) IS shipped but adds zero automation beyond UI tracking;
  Docker+SQLite infra unjustified.

Cites:
- SW-A merged PR #268 (Variant A REFERENCE, 28% match, 3 ADOPT-blockers)
- SW-B merged PR #267 (Variant B REJECT, ~5% match, no dir-watch capability)
- PR #269 follow-up (mechanical corrections, no verdict changes)
- DN-1=B-constrained input consumed in criterion 5 (mooted for Variant C
  which bypasses aif-handoff Reviewer entirely)
- Gate-4 admission re-sweep: PR #127+#128 touch packages/runtime/ only
  (no MCP/coordinator drift in 30-day window)

5 distinct DeepWiki probes + 2 WebSearches + cross-ref to SW-A/SW-B/PR #269
= 19+ evidence channels (T1 floor exceeded 3.8x). §1.7 forward+backward +
§self-application + T-trap walk per ai-laziness-traps.md §3.

Single output file under docs/meta-factory/research-patches/. No code, skill,
agent, install.sh, or .claude/rules/ modifications.

### §1.7 Forward-check applied

build-first-reuse-default.md §1 verdict ladder applied; BFR §3 6-layer search
performed (SSOT rows #27/#28/#29/#30/#43/#44/#46/#67/#80 reviewed at
prior-art-evaluations.md:95-148; DeepWiki >=5 probes; WebSearch >=2 phrasings;
own-stack sweep at .claude/skills/meta-orchestrator/SKILL.md:441 anti-scope +
:404+:429 SP requesting-code-review). no-paid-llm-in-ci.md §1 enforced (all
evidence via subscription-bundled DeepWiki/WebSearch + free gh CLI + bash).
reviewer-discipline.md §2 respected (DN-1=B-constrained consumed as fact, not
re-litigated; verdict is research finding against §8 STOP, not strategy
choice). ai-laziness-traps.md §3 active T-traps applied (T1, T3, T7, T11,
T12, T13, T15, T16, T17, T19, T20, T-AIF-BRIDGE-C MANDATORY BEFORE/AFTER
table at patch §4). Evidence: see patch §8 file:line citations.

### §1.7 Backward-check applied

SSOT #27/#28/#67 receive additive notes (additive-only; no verdict changes).
Original DEFER/DEFER/REJECT rationales reviewed at prior-art-evaluations.md:
95, 96, 135 — consistent with Sub-wave C findings (reinforce existing
classifications, do not re-litigate). No .claude/rules/* modified; no
.claude/skills/* modified; no agents/* modified; no packages/* modified;
no install.sh modified; no kickoff.md modified. Single output file in
docs/meta-factory/research-patches/. Scope strictly bounded to Variant C;
SW-A/SW-B/SW-B2/SW-D out of scope. T15 self-application confirmed in
patch §10. Memory not written (Sub-wave D synthesis is the natural
codification surface). Evidence: see patch §9 file:line citations.
artyhoo added a commit that referenced this pull request May 29, 2026
…verdict (DEFER-all-with-updated-triggers) (#281)

Cross-variant synthesis consolidating Sub-waves A/B/B'/C (merged PRs #267/#268/#275/#276). Per-variant verdicts re-derived independently per T15: A=REFERENCE (28%), B=REJECT (~5%), B'=REFERENCE conditional ADOPT VOCABULARY (~35%), C=REJECT (~22%). Cross-variant recommendation per kickoff §6.5: DEFER all variants with updated SSOT «Trigger to revisit» fields.

Includes 20-cell comparison table (4 variants × 5 criteria, file:line citation per cell), 4-row T16 problem-class match table, DN-1 (B-constrained) + DN-2 (B' is operative B-path) reflections, hook-discipline (Decision 10) analysis (no equipoise → recommend on clearly-better evidence; DECISION-NEEDED NOT warranted), consolidated SSOT additive notes for #27/#28/#30/#44/#67/#80, proposed new SSOT row #85 (next-after-#84 verified).

Prior-art: prior-art-evaluations.md#27, #28, #30, #44, #67, #80 (existing aif-handoff entries, additive-note-only — no verdict changes); SSOT row #85 PROPOSED, not landed.
artyhoo added a commit that referenced this pull request May 29, 2026
…DONE (#283)

Per umbrella's post-GO directive (Phase -1 Stage 3 close-review GO 2026-05-29 on
PR #281 SW-D synthesis), update wave-sequencing-plan.md §0 to reflect umbrella
closure:

- Row update: stale 🟡 «Sub-wave A + B R-phase done / synthesis pending» →
  ✅ DONE 2026-05-29 with verdict DEFER ALL. Full per-variant scoring
  (A REFERENCE 28% / B REJECT ~5% / B' REFERENCE 35% conditional / C REJECT 22%)
  + 3 ADOPT-blockers shared across A/B'/C + 6 evidence PRs (#267/#268/#269/
  #275/#276/#281) + SSOT additive notes status (proposed, not landed).
- Renamed row to canonical umbrella name (`aif-handoff-as-runtime-bridge`).
- «What actually remains» — remove the umbrella entry; add to deferred-decisions
  parenthetical at bottom of the line.
- Plan revision history — add this PR.

No verdict changes to any other §0 row. No edits outside §0. Broader 40+ UNTRACKED
PRs in §0 (recordkeeping debt per state.md Decision 2) stay deferred — out of
scope for this umbrella's closure per CLAUDE.md PR strategy.

Prior-art: skipped — recordkeeping update for completed umbrella, no capability
commit, no new dependency, no new file ≥50 LOC. Per CLAUDE.md capability-commit
definition, this is a docs-edit grade change.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant