[piyushmakhija28]

Closes item 3 of #296. Five modules, 538 lines, no importers. CLAUDE.md has described the package as "scheduled for removal in a follow-up cleanup" since the subpackage migration.

"No importers" was not the whole story, and checking changed what this PR does

All nine of the package's public node functions are also defined in orchestrator.py — the module create_flow_graph actually wires. So this was an extraction begun and never finished, not a component whose callers went away. Both copies were last touched by the same commit (06b0463, the Level/Step rename), which had been maintaining them in lockstep.

Comparing them function by function: five character-identical, four divergent. Most of the divergence is what a deeper package needs — parent.parent.parent instead of parent.parent, renamed module constants, different log tags. But two are documentation corrections that exist only in the orphaned copy, with the stale text in the live one:

	orphan (correct)	orchestrator.py (stale)
save_workflow_memory comment	~/.claude/memory/logs/sessions/	~/.claude/memory/sessions/ — contradicted the line directly beneath it, which writes "memory" / "logs" / "sessions"
optimize_context_after_level1	"Level 2"	"Level 3" ×3 — a level that stopped existing in 06b0463

Both are ported into orchestrator.py before the deletion, not alongside it. Deleting first and noticing afterwards is how a fix disappears with nothing recording that it existed.

The new enforcement test earned its place immediately

The Quick Info counts move 253 → 248 and 493 → 488. I did not spot that by hand — tests/test_doc_counts_match_the_tree.py, merged one commit earlier in #297, went red on the deletion and named both figures:

documentation counts disagree with the tree:
{'CLAUDE.md Quick Info: Total Python Files': {'says': (253, 493), 'actual': (248, 488)}}

First real change after that test landed, and it caught it.

Verified

orchestrator imports and create_flow_graph(hook_mode=True) builds a graph; all nine functions remain callable on it. Every test that touches the orchestrator passes (10 files). All six CI gates pass. Test collection is clean. No .py file anywhere still names helper_nodes.

Not verified locally: ruff is not on PATH in this shell, so the lint gate was left to pre-commit and CI. The changes are three comment lines plus deletions, so the exposure is small — but I did not run it myself and would rather say so.

Left alone deliberately

CHANGELOG.md's historical entry, docs/architecture/impact_map-l3-collapse-draft.md and docs/orchestration_prompt.md still name the package. The first two are point-in-time records; the third explicitly sequenced this removal as separate work, which is what happened. Rewriting them would falsify the history rather than correct it.

🤖 Generated with Claude Code