Assumption Testing

Every solution rests on assumptions. Test the riskiest ones first with the lightest method possible.

Preflight: Read target canvas file(s) before any Write/Edit

Hard rule. Before issuing Write or Edit against any .claude/canvas/*.yml, use the Read tool on that file in this session. Claude Code's Read-before-Write check requires the Read tool specifically — cat/head/grep via Bash do NOT satisfy it.

Edit vs Write — different cost profiles (verified 2026-05-14):

• Edit (exact-string replacement): Read with limit: 1 satisfies the check at ~50 tokens. State-tracking is per-file, not per-byte — subsequent Edit calls work anywhere in the file. Use this for partial updates against large canvas files (e.g., purpose.yml at 800+ lines).
• Write (full replacement): do a full Read first. Write obliterates the file; you should see what you're about to replace. The limit:1 shortcut is not appropriate here.

ID-bearing entries — scan the ID space before assigning (added 2026-05-15, v0.23.19): When adding a new component, opportunity, solution, or any other ID-bearing entry to a canvas file, run a Bash grep first to confirm the next ID in your prefix sequence is actually free:

grep "^  - id: <prefix>-" .claude/canvas/<file>.yml | sort -u

Replace <prefix> with the canvas's ID prefix (comp for landscape, opp for opportunities, sol for solutions, ht for human-tasks, etc.). Then pick the next free integer. validate_canvas.py has a duplicate-ID check (lines 230-239) that catches the failure on CI, but a duplicate can persist in the working tree for days if CI isn't run between edit and discovery — see roadmap-repo corrections.md 2026-05-15 "Duplicate canvas ID created in landscape.yml" for the worked example.

Original failure mode: anti-pattern #7 instance #5, 2026-05-09 — agent conflated Bash head with the Read tool, lost ~14k tokens to a Write-fail → remedial-full-Read → re-Write loop. The limit:1 discipline (graduated 2026-05-14, v0.23.18) prevents the second-order cost where the agent correctly follows the rule but full-Reads every time. The ID-scan discipline (graduated 2026-05-15, v0.23.19) prevents the related class where the agent reads enough of the file to satisfy the Edit check but not enough to see existing ID assignments — kin to anti-pattern #8 (Stale State Read).

If this skill writes to multiple canvas files, register each one first (limit:1 for Edit-only paths; full Read for Write paths) AND ID-scan any prefix you intend to assign.

See CLAUDE.md Canvas writes — Read before Write for the canonical rule.

Assumption Types (Torres / Cagan)

Step 1: Map Assumptions

For the target solution, list ALL assumptions. Be honest -- most "obvious" things are actually assumptions.

Couple the test to open canvas gaps (per engine/canvas-guidance.yml#learning_target_coupling): before finalizing the assumption list, scan the canvas for entries already waiting on evidence — ON HOLD / RE-GATED action flags, in-progress human-tasks naming a MISSING SIGNAL, low-confidence entries with an un-validated assumption. If this test touches any of them, add that gap to the list explicitly and tag it [target → <file>#<anchor>] so /mycelium:log-evidence routes the result back. A test that retires no open gap spends scarce feedback capacity without advancing the canvas. NUDGE-tier — zero-target tests are allowed, but make it a choice.

Step 2: Prioritize (2x2 Matrix)

Plot each assumption on:

• X-axis: How much evidence do we have? (low to high)
• Y-axis: How important is this to the solution's success? (low to high)

Test first: High importance + Low evidence (top-left quadrant)

Step 3: Choose the Lightest Test

Organized by Gilad's AFTER model (Assessment → Fact-Finding → Tests → Experiments → Release Results). Always start from the top and pick the lightest test that produces meaningful signal. Don't build a prototype when a survey would suffice.

Assessment (internal, cheapest — hours)

Fact-Finding (external evidence — hours to days)

Tests (controlled artifacts — days to weeks)

Experiments (statistical comparisons — weeks)

Release Results (staged release — weeks)

Source: Gilad (AFTER model, Evidence-Guided / Testing Product Ideas Handbook). 28 techniques across 5 stages, ordered by cost and confidence.

Session-counter primitive (for shadow logs / longitudinal tests)

Tests in