Skip to content

RF โ€” TFW-42 / Phase A: Conventions & Templates

Date: 2026-04-30 Author: Executor (Antigravity) Status: ๐ŸŸข RF โ€” Complete Parent HL: HL-TFW-42 TS: TS Phase A


1. What Was Done

New Files

File Description
.tfw/templates/research/1_briefing.md Renamed stage template (was briefing.md)
.tfw/templates/research/2_gather.md Renamed stage template (was gather.md)
.tfw/templates/research/3_extract.md Renamed stage template (was extract.md)
.tfw/templates/research/4_challenge.md Renamed stage template (was challenge.md)

Modified Files

File Changes
.tfw/conventions.md ยง4 filename table: Phase{X} โ†’ phase-{x} in all 5 Phase rows. ยง4 Research subfolder: complete rewrite โ€” unified research/ container, research/iterN/ pattern, numbered stage files, co-located RES.md. ยง4 iterations.yaml: moved from task root to research/, enriched schema with agent + sources optional fields, updated res_file paths. ยง4 new Agent selection guidance subsection (5-row capability table). ยง4 Review subfolder: updated parallel reference from research/gather.md to research/iterN/1_briefing.md. ยง4 Multi-phase folder structure: PhaseA/ โ†’ phase-a/, updated all artifact filenames in tree. Fixed section ordering (Research โ†’ iterations.yaml โ†’ Agent guidance โ†’ Review โ†’ Multi-phase).
.tfw/templates/RES.md Line 15: research/briefing.md โ†’ 1_briefing.md in iteration folder

Deleted Files

File Reason
.tfw/templates/research/briefing.md Renamed to 1_briefing.md
.tfw/templates/research/gather.md Renamed to 2_gather.md
.tfw/templates/research/extract.md Renamed to 3_extract.md
.tfw/templates/research/challenge.md Renamed to 4_challenge.md

2. Key Decisions

  1. Section reordering in ยง4. The TS assumed iterations.yaml block was after Review subfolder, but actually Review subfolder (line 148) was between Research subfolder (131) and iterations.yaml (152). Restructured to logical grouping: Research subfolder โ†’ Multi-iteration โ†’ iterations.yaml โ†’ Agent guidance โ†’ Review subfolder โ†’ Multi-phase. Coordinator approved this in ONB ยง6.

  2. RES template briefing path made relative. Since RES.md now lives inside research/iterN/ alongside its stage files, changed the reference from absolute research/briefing.md to relative 1_briefing.md in iteration folder.

  3. Multi-phase tree updated. Replaced RES__PROJ-5__query_redesign.md โ† Master RES (if any) with research/ โ† Master research (if any) to align with the new container convention.

3. Acceptance Criteria

  • [x] AC-1: Research folder structure โ€” research/ is single container, research/iterations.yaml documented, research/iterN/ pattern used, research/iterN/RES.md co-located, trace rule preserved, table shows iter1/iter2/iterN paths
  • [x] AC-2: iterations.yaml schema โ€” agent: field present (commented-out, marked optional), sources: field present (commented-out, marked optional), res_file paths use research/iterN/RES.md, no brief/notes/depends_on fields, text explains "traceability, not dispatch"
  • [x] AC-3: Numbered stage files โ€” 1_briefing.md, 2_gather.md, 3_extract.md, 4_challenge.md exist in templates/research/; old files deleted; conventions.md ยง4 lists numbered filenames
  • [x] AC-4: Phase folder naming โ€” folder example shows phase-a/, phase-b/; filename table uses phase-{x} format and phase-a example; 0 matches for PhaseA in ยง4
  • [x] AC-5: Schema backward compatibility โ€” all 9 original fields preserved (task_id, title, min_iterations, max_iterations, iterations[] with number, focus, hypotheses, status, res_file); agent and sources are additive
  • [x] AC-6: Agent selection guidance โ€” subsection titled "Agent selection guidance", 5-row capability table (web research, code audit, infra recon, architecture synthesis, data analysis), columns: Research Activity / Key Capability / When to consider, no tool brand names, footer: "guidance, not prescription"

4. Verification

  • Lint: N/A (markdown documentation changes only)
  • Tests: N/A (no test infrastructure for conventions)
  • Verify:
  • ls .tfw/templates/research/ โ†’ 4 files with numeric prefixes, 0 without โœ…
  • Grep PhaseA in conventions.md โ†’ 0 matches โœ…
  • Grep researchN/, research2/, research3/ in conventions.md โ†’ 0 matches โœ…
  • Grep depends_on, brief:, notes: in conventions.md โ†’ 0 matches โœ…
  • Visual verification of all updated sections โ€” confirmed correct โœ…

5. Observations (out-of-scope, not modified)

# File Line(s) Type Description
1 .tfw/compilable_contract.md 56, 78 naming References PhaseA/ in resolution rules โ€” needs update when conventions adopt kebab-case. Not in Phase A scope
2 .tfw/workflows/handoff.md 140-141 naming Multi-Phase Task Flow example uses HL__PhaseA, TS__PhaseA, RF__PhaseA โ€” needs Phase B/C update
3 .tfw/workflows/plan.md 134-136 naming Multi-phase structure example uses PhaseA/, HL__PhaseA__, TS__PhaseA__ โ€” needs Phase B update
4 .tfw/workflows/research/base.md 53 naming References research/briefing.md and researchN/briefing.md โ€” needs Phase B update to use research/iterN/1_briefing.md
5 .tfw/CHANGELOG.md 284 naming Historical reference to PhaseA/, PhaseB/ in v0.6.0 โ€” historical, no action needed

6. Fact Candidates

No fact candidates.

7. Strategic Insights (Execution)

No strategic insights.

8. Diagrams

No diagrams.


RF โ€” TFW-42 / Phase A: Conventions & Templates | 2026-04-30