ace-test-runner-e2e 0.38.11 → 0.40.1

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 1a1e81b2b077a6bca7e75e1572743a31f20abe1ef5ebcb69ea82c7f55e95fd4b
-  data.tar.gz: e1791696e6cbbb58decab800387e005ee22f134734be2736d5c00f109273dd57
+  metadata.gz: ae94a3ebd8b4ed697d8b3b5a705030236659ab9b8e34fa4e5baa421a22d1b781
+  data.tar.gz: 8d65447a2174a8fe614d0fce6a8ab2503d39b06e76584aa7af4ba9cedaf3f8bc
 SHA512:
-  metadata.gz: 143efde4ad09db543ff0865da3de1a94343c278a64c12f55f1786718725846b5cec565e144687b8c9d16bbc15736bf73abb34b30b49fa746a56ed522978e6434
-  data.tar.gz: b166bec29e9f10d0eff3b692d6526c22e0960252fb1a9d6d125c646b47a14137472775f73a9120365ebb608d7f2ccc38f88b4694e1b144165cf1a2994512ed2e
+  metadata.gz: 3889a846fd3631330728fe5e144259328f08ebd7ee35d4d8f5358fb149f85ad1dc4071132d1db283bb402f1cbed16655f1e790a644e40508e1d104b4fb24e0f5
+  data.tar.gz: c68c945a12f8ab86c23ad3714584afbc9e507a7cc8d91b089c99048e43df579adae9f407e769e9669e17910c5853c0919b81846fb278b5d72722220b29187101

data/CHANGELOG.md

@@ -7,6 +7,61 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ## [Unreleased]
 
+## [0.40.1] - 2026-04-24
+
+### Fixed
+- Removed suite-specific wording from the single-command `ace-test-e2e` help/output path so the `RunTest` CLI stays scoped to the single-command surface while preserving prune-artifact guidance.
+
+## [0.40.0] - 2026-04-24
+
+### Changed
+- Added `--[no-]prune-artifacts` to `ace-test-e2e` and `ace-test-e2e-suite` so operators can clear stale `.ace-local/test-e2e` run artifacts before execution while preserving suite reports and the shared `runtime-cache/`.
+
+## [0.39.1] - 2026-04-24
+
+### Fixed
+- Resolved suite-shared runtime reuse in child `ace-test-e2e` subprocesses by honoring inherited `ACE_E2E_SHARED_RUNTIME_ROOT` from the process environment instead of rebuilding sandbox-local runtimes after prewarming.
+
+## [0.39.0] - 2026-04-24
+
+### Changed
+- Added `--[no-]retry-failures-once` for full-suite reruns, including flaky-recovery reporting when a failed first pass succeeds on the retry pass.
+- Reused a suite-shared E2E runtime cache under `.ace-local/test-e2e/runtime-cache/` so parallel sandbox workers stop rebuilding the same Bundler environment and native extensions for every scenario.
+
+## [0.38.17] - 2026-04-24
+
+### Fixed
+- Detected fixture-commit setup flows across the full setup sequence instead of only single-step `git add && git commit` commands, restoring support-path git excludes for split-step fixture repositories.
+
+## [0.38.16] - 2026-04-23
+
+### Fixed
+- Enforced runner-owned verifier artifact contracts in scenario loading, expanded grouped `.stdout` / `.stderr` / `.exit` shorthand, and rejected verifier-only or wildcard artifact declarations that previously let retained E2E drift slip through.
+
+### Technical
+- Updated E2E guides, templates, and create/review/plan/rewrite/fix workflows to distinguish `public-surface` versus `retained-contract` TCs and require explicit downstream retained-E2E sweeps after public contract changes.
+
+## [0.38.15] - 2026-04-23
+
+### Fixed
+- Passed declared artifact contracts directly into runner prompts, added one bounded runner repair pass when required captures are still missing, and persisted repair metadata so missing-artifact E2E failures can recover before verifier judgment.
+
+## [0.38.14] - 2026-04-23
+
+### Fixed
+- Limited deterministic sandbox git excludes to setup-commit scenarios so copied package trees remain visible to ignore-aware tools while fixture-repo support paths stay unstaged.
+
+## [0.38.13] - 2026-04-23
+
+### Fixed
+- Enabled role-based verifier fallback in pipeline execution so successful runner phases still produce verifier results when the first verifier provider is unavailable.
+- Seeded deterministic sandbox git excludes for copied package trees and fixture-commit support paths so setup-time `git add -A` no longer stages runner support files or copied package content into fixture repositories.
+
+## [0.38.12] - 2026-04-23
+
+### Changed
+- Updated default ACE sandbox bootstrap to use `ace-config sync ace-llm-providers-cli` before `ace-handbook sync`, matching the renamed config sync command and minimal quick-start config requirement.
+
 ## [0.38.11] - 2026-04-20
 
 ### Fixed

data/handbook/guides/e2e-testing.g.md

@@ -14,6 +14,10 @@ ace-docs:
 E2E tests are executed by an AI agent and reserved for behaviors that require real CLI execution, real tools, and real filesystem side effects.
 They must also answer a user-journey question: can a user do the job from the tool's public surface, and how much friction does that journey have?
 
+In practice, ACE uses two valid TC styles:
+- **Public-surface TCs** — prove a user job from docs/usage/`--help` and the CLI itself.
+- **Retained-contract TCs** — pin a previously fragile integrated behavior with deterministic, explicitly declared evidence.
+
 ## Canonical Conventions
 
 - CLI split:
@@ -33,7 +37,7 @@ They must also answer a user-journey question: can a user do the job from the to
 
 - Runner is **execution-only**:
   - perform user-like CLI actions in sandbox
-  - produce only final outcome evidence under `results/tc/{NN}/`
+  - produce only declared outcome evidence under `results/tc/{NN}/`
   - return final runner observations through the harness contract
   - do not issue PASS/FAIL verdicts
   - do not perform verifier-style assertion/classification
@@ -46,6 +50,11 @@ They must also answer a user-journey question: can a user do the job from the to
     2. runner observations
     3. explicit TC artifacts that are true product outcomes
     4. debug captures (`stdout`, `stderr`, `*.exit`, metadata) only as fallback
+- Artifact contract ownership:
+  - runner instructions and `scenario.yml` setup/layout declare verifier-visible artifact paths
+  - verifier consumes that contract; it does not create new required artifact paths
+  - grouped shorthand such as ``results/tc/02/help.stdout`, `.stderr`, `.exit`` counts as an exact declaration of all three files
+  - wildcard artifact paths such as `results/tc/02/output.*` are not valid declarations
 - Setup ownership:
   - sandbox preparation belongs to `scenario.yml` `setup:` + `fixtures/`
   - TC runner files must not define independent environment setup procedures
@@ -77,6 +86,21 @@ When an E2E failure shows that a valid user job is not discoverable from docs, u
 docs/help drift. Failure analysis must record the stale or missing public surface and the exact docs/help target to
 update instead of teaching the runner a workaround.
 
+## TC Style Selection
+
+Use **public-surface** style when the goal is a real user journey and the primary oracle should stay on user-visible behavior.
+
+Use **retained-contract** style when the integrated behavior matters but final sandbox state alone is not enough. In that case, small declared supporting captures are valid, for example:
+- `.stdout`, `.stderr`, `.exit`
+- `command.txt`
+- `path-check.txt`
+- `artifact-check.txt`
+
+Even retained-contract TCs must not rely on:
+- verifier-only artifact declarations
+- wildcard artifact paths
+- reflections, PASS/FAIL summaries, or verifier-facing manifests under `results/`
+
 ## Cost and Scope
 
 - Keep scenarios small and coherent.
@@ -128,12 +152,16 @@ This prevents duplicate assertions across test layers.
 - Keep runner goals aligned with the public user path; if the runner needs a workaround, surface that as friction rather than teaching the workaround.
 - Keep verifier expectations impact-first, then artifacts, then debug fallback.
 - Preserve strict TC pairing (`runner` + `verify`).
-- Keep `results/tc/{NN}/` for outcome artifacts only.
-- Do not instruct runners to create helper YAML, path files, command files, or reflections in `results/`.
+- Keep `results/tc/{NN}/` for declared verifier-dependent evidence only.
+- Declare every verifier-dependent file path in runner instructions or scenario setup. Do not rely on verifier-only path references.
+- Allow small supporting captures only when they are explicitly declared and materially improve confidence.
+- Do not use wildcard artifact paths.
+- Do not instruct runners to create reflections, PASS/FAIL summaries, verifier-facing manifests, or ad hoc temp inputs in `results/`.
 - Do not judge success from runner-authored summaries when final sandbox state can prove the goal directly.
 - Use runner observations only to explain ambiguity or missing side effects, not to replace missing end-state evidence.
 - Treat any workaround noted in runner observations as a product/docs/help or scenario-design smell that must be fixed, not preserved.
 - Avoid hidden dependencies between TCs unless explicitly intended.
+- For `--watch` or other live-output commands, use a bounded-session pattern with explicit termination behavior and captured exit codes.
 
 ## Execution Artifacts
 
@@ -150,9 +178,13 @@ Before approving new/updated E2E tests:
 - [ ] `runner.yml.md` and `verifier.yml.md` exist
 - [ ] Every TC has both `.runner.md` and `.verify.md`
 - [ ] Artifacts are scoped to `results/tc/{NN}/`
+- [ ] Every verifier-dependent artifact path is declared by runner/setup
+- [ ] No verifier depends on wildcard or verifier-only artifact paths
 - [ ] Verifier primary oracle is final sandbox state or real product output, not helper artifacts
 - [ ] Runner observations are the only non-filesystem secondary evidence source
+- [ ] TC style is explicit in the review (`public-surface` or `retained-contract`)
 - [ ] Scenario can be completed from docs/usage/`--help` without hidden recipes or workaround instructions
+- [ ] Any internal-detail assertion is part of the public contract or justified as retained-contract evidence
 - [ ] Any friction/workaround found during review is treated as a gap, not as a runner script opportunity
 - [ ] Failure analysis records docs/help drift from failed public user paths, or explicitly records `None`
 - [ ] Value-gate metadata is present (`e2e-justification`, `unit-coverage-reviewed`, `cost-tier`)

data/handbook/guides/scenario-yml-reference.g.md

@@ -46,7 +46,7 @@ Example: `ace-lint/test/e2e/TS-LINT-001-lint-pipeline/scenario.yml`
 |-------|------|---------|-------------|
 | `priority` | string | `medium` | Test priority: `high`, `medium`, `low` |
 | `tool-under-test` | string | — | Primary command/tool validated |
-| `sandbox-layout` | object | `{}` | Outcome-path hints used to precreate directories and guide verification |
+| `sandbox-layout` | object | `{}` | Directory-level outcome hints used to precreate `results/tc/*` paths and guide verification |
 | `duration` | string | — | Estimated duration (e.g., `~15min`) |
 | `timeout` | integer | — | Optional per-scenario execution timeout in seconds |
 | `automation-candidate` | boolean | `false` | Whether test is automatable |
@@ -73,7 +73,10 @@ Pairing rule:
 Artifact layout conventions:
 - canonical: `results/tc/{NN}/`
 - avoid non-TC-scoped result folders
-- keep only real outcome artifacts under `results/tc/{NN}/`; runner observations live in harness reports, not sandbox helper files
+- keep only declared verifier-dependent evidence under `results/tc/{NN}/`; runner observations live in harness reports, not sandbox helper files
+- file-level verifier checks must be declared by the runner; `sandbox-layout` does not replace exact file declarations
+- grouped shorthand such as ``results/tc/01/help.stdout`, `.stderr`, `.exit`` is valid for exact sibling captures
+- wildcard artifact paths are not supported
 - absence of a declared path is debug context, not a standalone failure reason
 
 Canonical summary report fields:
@@ -85,7 +88,8 @@ Canonical summary report fields:
 Role contract:
 - `runner.yml.md` + `TC-*.runner.md` are execution-only.
 - `verifier.yml.md` + `TC-*.verify.md` are verification-only with impact-first checks.
-- Goal-style scenarios should be solvable from the public surface (docs/usage/`--help` + tool under test) without hidden recipes or workaround instructions.
+- Public-surface TCs should be solvable from the public surface (docs/usage/`--help` + tool under test) without hidden recipes or workaround instructions.
+- Retained-contract TCs may keep small declared supporting captures when they materially improve confidence.
 
 ## `requires` Object
 
@@ -130,6 +134,7 @@ setup:
 Setup rules:
 - Setup is fail-fast. Do not hide setup failures with `|| true`.
 - Setup belongs in `scenario.yml` and fixtures, not in TC runner instructions.
+- Use setup to create prerequisite state, not verifier-facing helper files under `results/`.
 - If setup fails (for example, missing `mise trust` support), stop scenario execution and report infrastructure failure.
 
 ## Complete Example

data/handbook/guides/tc-authoring.g.md

@@ -32,6 +32,11 @@ Inline `.tc.md` and frontmatter `mode` values are no longer supported.
 - TC outcome artifacts write to `results/tc/{NN}/`
 - Summary counters use `tcs-passed`, `tcs-failed`, and `tcs-total`
 
+## TC Styles
+
+- **Public-surface**: prove a documented user job from docs/usage/`--help` and the CLI.
+- **Retained-contract**: pin an integrated behavior with deterministic, explicitly declared supporting evidence when end-state checks alone are insufficient.
+
 ## File Naming
 
 - `TC-{NNN}` — test case number (e.g., TC-001)
@@ -82,8 +87,9 @@ Run `ace-lint` and produce report artifacts for a valid file.
 ## Constraints
 
 - Use only sandbox paths
-- Keep only final outcome evidence under `results/tc/01/`
-- Do not place helper inputs, manifests, command transcripts, or reflections under `results/tc/01/`
+- Keep only declared verifier-dependent evidence under `results/tc/01/`
+- Declare exact paths for any verifier-dependent captures, for example ``results/tc/01/help.stdout`, `.stderr`, `.exit``
+- Do not place helper inputs, manifests, PASS/FAIL summaries, or reflections under `results/tc/01/`
 - Execute actions only; do not assign PASS/FAIL or final verdicts
 ```
 
@@ -122,14 +128,19 @@ Pass only when all expectations are satisfied by on-disk evidence.
 
 - Keep each TC focused on one coherent behavior path.
 - Ensure goal numbers and TC numbers remain aligned (`TC-001` -> Goal 1).
+- Choose the TC style up front: `public-surface` or `retained-contract`.
 - Keep runner files execution-only and verifier files verdict-only.
 - Make verifier expectations deterministic with impact-first ordering.
-- Keep `results/tc/{NN}/` for outcome artifacts only.
+- Keep `results/tc/{NN}/` for declared verifier-dependent evidence only.
+- Declare every verifier-dependent path in the runner or setup. Do not rely on verifier-only references.
+- Grouped capture shorthand is valid only for exact sibling files, for example ``foo.stdout`, `.stderr`, `.exit``.
+- Do not use wildcard artifact paths.
 - Use harness-provided runner observations as the only non-filesystem secondary evidence source.
 - Prefer final sandbox state and real product output over raw debug captures.
-- Do not ask the runner to write setup inputs, audit manifests, or final reflections for the verifier.
+- Do not ask the runner to write setup inputs, audit manifests, verifier-facing summaries, or final reflections for the verifier.
 - Do not teach the runner hidden recipes or workaround sequences; if the path is not discoverable from docs/usage/`--help`, the TC is wrong or the public surface needs improvement.
 - Use runner observations to record friction and workaround pressure, not to normalize it.
+- For watch/live-output flows, use a bounded-session pattern with explicit shutdown and captured exit code.
 - Record why each scenario remains E2E via `e2e-justification` and `unit-coverage-reviewed` in `scenario.yml`.
 
 ## Related

data/handbook/templates/tc-file.template.md

@@ -22,7 +22,9 @@ ace-docs:
 
 - Use only declared scenario tools (`ace-*` and explicit exceptions)
 - Keep only product outcomes or essential command captures under `results/tc/{NN}/`
-- Do not write helper inputs, reflections, manifests, or temp files under `results/tc/{NN}/`
+- Declare every verifier-dependent path explicitly in the runner or scenario setup
+- Grouped capture shorthand such as ``results/tc/{NN}/cmd.stdout`, `.stderr`, `.exit`` is allowed for exact sibling files
+- Do not write helper inputs, reflections, PASS/FAIL summaries, manifests, or temp files under `results/tc/{NN}/`
 - Do not write outside sandbox
 - Execute actions only; do not assign PASS/FAIL in runner file
 - Follow the public user path from docs/usage/`--help`; do not embed hidden recipes or workaround branches in the TC
@@ -51,5 +53,5 @@ Companion verifier file (`TC-{NNN}-{slug}.verify.md`) example:
 
 ## Verdict
 
-- Pass when the public path works from sandbox evidence. Missing helper artifacts alone should not fail the goal.
+- Pass when the public path or retained contract is satisfied from sandbox evidence. Undeclared helper artifacts alone should not fail the goal.
 -->

data/handbook/workflow-instructions/e2e/create.wf.md

@@ -41,7 +41,10 @@ This workflow guides an agent through creating a new E2E test scenario.
 ## Authoring Contract
 
 - Runner files (`runner.yml.md`, `TC-*.runner.md`) are execution-only.
-- Goal-style TCs must prove two things:
+- Every TC must be authored as one of:
+  - **public-surface** — a user job from docs/usage/`--help` and the CLI
+  - **retained-contract** — a deterministic integrated regression check with declared supporting evidence
+- Goal-style/public-surface TCs must prove two things:
   - the tool works
   - a user can do the job from the public surface (`README`, usage docs, `--help`, and the CLI itself) without hidden recipes or workarounds
 - Verifier files (`verifier.yml.md`, `TC-*.verify.md`) are verdict-only with impact-first evidence order:
@@ -52,7 +55,10 @@ This workflow guides an agent through creating a new E2E test scenario.
   4. debug captures as fallback
 
 - Setup belongs to `scenario.yml` `setup:` and fixtures; do not duplicate setup in runner TC instructions.
-- Keep `results/tc/{NN}/` for real outcome artifacts only; do not ask the runner to write helper YAML, path files, command files, reflections, or verifier-facing manifests there.
+- Keep `results/tc/{NN}/` for declared verifier-dependent evidence only.
+- Declare every verifier-dependent path in the runner or setup. Grouped shorthand such as ``foo.stdout`, `.stderr`, `.exit`` is allowed for exact sibling captures.
+- Do not use wildcard artifact paths.
+- Do not ask the runner to write reflections, verifier-facing manifests, or undeclared helper files there.
 - Do not encode hidden command recipes, fallback detours, or workaround sequences in runner TC files. If the job cannot be done from the public surface, treat that as a product/docs/help gap or remove/narrow the TC.
 
 ## Workflow Steps
@@ -248,9 +254,12 @@ Rules:
 - `existence-only` is never valid for KEEP/ADD. Use it only for SKIP rows with explicit unit-test replacement.
 - `helper-artifact-driven` is never valid for KEEP/ADD when final sandbox state could prove the goal directly.
 - `hidden-recipe-driven` and `workaround-driven` are never valid for KEEP/ADD.
+- Every verifier-dependent artifact must be declared by runner/setup; verifier-only references are invalid.
+- Wildcard artifact paths are never valid for KEEP/ADD.
 - `SKIP` rows must include replacement unit-test evidence.
 - Non-skipped rows must identify the primary oracle for the TC: final sandbox state, real product output, or debug fallback.
 - Non-skipped rows must state why the job is achievable from the public surface without hidden recipes.
+- Non-skipped rows must identify TC style: `public-surface` or `retained-contract`.
 - At least one `unit tests reviewed` path is required for every row.
 - The scenario-level `unit-coverage-reviewed` field must include the union of all referenced unit test files.
 
@@ -267,6 +276,7 @@ Rules:
 - No TC may be created without a row in this table.
 - If decision is `SKIP`, include the unit-test evidence that replaces it.
 - If the public-surface path is missing or workaround-driven, the TC must be `SKIP` or explicitly planned as a product/docs/help improvement before creation.
+- If the TC uses live refresh or watch behavior, include a bounded-session capture plan with explicit shutdown behavior and exit-code expectations.
 - At least one `unit tests reviewed` path is required for each row.
 - The scenario-level `unit-coverage-reviewed` field must include the union of all referenced unit test files.
 
@@ -301,7 +311,7 @@ If a context description was provided, enhance the test with:
 - Write runner goals as user outcomes, not “create a report” chores for the verifier
 - Check specific exit codes for error commands (not just "non-zero")
 - Make final sandbox state or real product output the primary oracle whenever possible
-- Do not require runner-authored helper files under `results/tc/{NN}/`
+- Do not require undeclared or verifier-facing helper files under `results/tc/{NN}/`
 - Add at least one behavioral/content assertion when CLI output itself is part of the outcome being tested
 
 **SHOULD (strongly recommended):**

data/handbook/workflow-instructions/e2e/fix.wf.md

@@ -110,6 +110,7 @@ Apply fixes in this order:
 - Preserve role split: runner is execution-only, verifier is impact-first verdict
 - Keep implementation unchanged unless analysis is revised
 - Remove hidden recipes, workaround branches, and unsupported internal-detail checks from goal-style TCs
+- Repair undeclared or wildcard artifact contracts before weakening product assertions
 
 4. Rerun the selected failing scope after each fix
 
@@ -150,6 +151,12 @@ ace-test-e2e ace-bundle TS-BUNDLE-001
 - Keep one active scenario/TC at a time
 - Preserve cost-conscious rerun discipline
 
+6a. If the fix changes a public contract, run a downstream retained-E2E sweep
+
+- Trigger this sweep when the fix changes status words, JSON keys, command shapes, lifecycle semantics, or ownership/state semantics
+- Grep impacted scenarios and downstream consumers before concluding the fix
+- Update retained runner/verifier contracts in the same change set whenever feasible
+
 7. Run a final explicit failing-scenario checkpoint before concluding the fix session
 
 After the currently targeted failures are addressed, require one final:
@@ -179,6 +186,18 @@ Analysis Source: reused existing analysis | generated via `wfi://e2e/analyze-fai
 | ... | ... | ... | ... | pass/fail |
 ```
 
+Also include:
+
+```markdown
+## Fix Classification Totals
+
+| Bucket | Count |
+|---|---|
+| Product bug | {n} |
+| Harness bug | {n} |
+| Retained test/spec drift | {n} |
+```
+
 If the analysis reported docs/help drift, include:
 
 ```markdown

data/handbook/workflow-instructions/e2e/plan-changes.wf.md

@@ -56,6 +56,15 @@ Build a change inventory:
 - **Removed features** — deleted files or deprecated modules
 - **Unchanged features** — stable code with no recent modifications
 
+Before classifying TCs, also check whether the package change alters a public contract that downstream retained E2E tests commonly pin:
+- status words
+- JSON keys or output schema
+- CLI command/flag shapes
+- lifecycle semantics
+- ownership/state semantics
+
+If yes, add an explicit downstream retained-E2E sweep list to the plan instead of limiting scope to the package under edit.
+
 ### 3. Classify Each Existing TC
 
 For each TC listed in the coverage matrix, assign exactly one classification:
@@ -80,6 +89,7 @@ For REMOVE due to overlap, replacement evidence is mandatory:
 - TC scope is too broad (should be narrowed to only E2E-exclusive aspects)
 - TC scope is too narrow (missing assertions for related behavior in same CLI invocation)
 - TC has structure issues flagged in the review
+- TC has undeclared or wildcard verifier-dependent artifact paths
 - TC is hidden-recipe-driven or workaround-driven but the underlying user job should still be supported by the public surface after scenario/docs/help correction
 
 **CONSOLIDATE** — The TC should merge with another TC. Criteria (any one is sufficient):
@@ -193,6 +203,12 @@ Format the complete change plan:
 |--------|--------|-----|
 | Update docs/help/CLI | {package/path} | {job is valid but current public surface is too weak for the E2E path} |
 
+### Downstream Retained-E2E Sweep ({n} actions)
+
+| Scenario | Trigger | Change Needed |
+|----------|---------|---------------|
+| {scenario-id} | {renamed key / lifecycle shift / command-shape change} | {update retained verifier/runner contract} |
+
 ### CONSOLIDATE ({n} TCs → {n} TCs)
 
 | Source TCs | Target TC | Merged Assertions |

data/handbook/workflow-instructions/e2e/review.wf.md

@@ -14,7 +14,8 @@ This workflow performs deep exploration of a package to produce a **coverage mat
 During review, treat the runner/verifier split as a first-class quality check:
 - Runner must be execution-only (no verdict language).
 - Verifier must be impact-first (sandbox impact before runner observations and debug).
-- `results/tc/{NN}/` must not be used for helper inputs or verifier-feeding helper reports.
+- `results/tc/{NN}/` must contain only declared verifier-dependent evidence.
+- Every verifier-dependent artifact path must be declared by runner/setup; verifier-only or wildcard references are contract drift.
 - Goal-style TCs must also pass the public-surface check: the runner should be able to do the job from docs/usage/`--help` and the tool under test, without hidden recipes or workarounds.
 
 **Pipeline position:** Stage 1 of 3 (Explore)
@@ -117,6 +118,9 @@ find {PACKAGE}/test/e2e -name "scenario.yml" -path "*/TS-*" 2>/dev/null | sort
   - `tags`, `cost-tier`, `e2e-justification`, `unit-coverage-reviewed`
   - `last-verified`, `verified-by`
 - Extract the objective (what the TC verifies)
+- Record TC style:
+  - `public-surface`
+  - `retained-contract`
 - Record the TC's primary oracle:
   - final sandbox state / real product output
   - runner observations as supporting context
@@ -134,15 +138,15 @@ find {PACKAGE}/test/e2e -name "scenario.yml" -path "*/TS-*" 2>/dev/null | sort
 - Mark TC evidence status:
   - `complete` when `e2e-justification` is present, the verifier is end-state-first, and `unit-coverage-reviewed` has at least one path
   - `missing` otherwise
-  - `at-risk` when evidence is existence-only, helper-artifact-driven, duplicate command invocations are detected, or the TC is hidden-recipe/workaround-driven
+  - `at-risk` when evidence is existence-only, helper-artifact-driven, duplicate command invocations are detected, the TC is hidden-recipe/workaround-driven, or verifier-dependent artifacts are undeclared
 
 If `--scope` was provided, filter to only the specified scenario.
 
 Build an E2E test map:
 
-| TC ID | Title | Command Invocations | Feature Tested | Primary Oracle | Public Surface Fit | Friction | Tags | Cost Tier | E2E Justification | Unit Coverage Reviewed | Evidence | False-Positive Risk |
-|-------|-------|-------------|----------------|----------------|--------------------|----------|------|-----------|-------------------|------------------------|----------|---------------------|
-| {id} | {title} | {command list} | {feature} | {state / output / observations+fallback} | {valid/hidden-recipe/workaround/unsupported-detail} | {low/medium/high} | {tags} | {tier} | {reason or "(missing)"} | {files or "(missing)"} | {complete/missing/at-risk} | {low/medium/high} |
+| TC ID | Style | Title | Command Invocations | Feature Tested | Primary Oracle | Public Surface Fit | Artifact Contract | Friction | Tags | Cost Tier | E2E Justification | Unit Coverage Reviewed | Evidence | False-Positive Risk |
+|-------|-------|-------|-------------|----------------|----------------|--------------------|-------------------|----------|------|-----------|-------------------|------------------------|----------|---------------------|
+| {id} | {public-surface/retained-contract} | {title} | {command list} | {feature} | {state / output / observations+fallback} | {valid/hidden-recipe/workaround/unsupported-detail} | {declared/undeclared/wildcard} | {low/medium/high} | {tags} | {tier} | {reason or "(missing)"} | {files or "(missing)"} | {complete/missing/at-risk} | {low/medium/high} |
 
 ### 5. Build Coverage Matrix
 
@@ -214,12 +218,12 @@ TCs that may fail the E2E Value Gate (unit tests cover the same behavior or high
 
 ### E2E Decision Record Coverage
 
-| TC ID | Evidence Status | Public Surface Fit | Friction | Missing Fields / Contract Drift |
-|-------|------------------|--------------------|----------|-------------------------------|
-| {id} | complete | valid | low | none |
-| {id} | missing | hidden-recipe-driven | high | e2e-justification, unit-coverage-reviewed, end-state oracle |
+| TC ID | Style | Evidence Status | Public Surface Fit | Artifact Contract | Friction | Missing Fields / Contract Drift |
+|-------|-------|------------------|--------------------|-------------------|----------|-------------------------------|
+| {id} | public-surface | complete | valid | declared | low | none |
+| {id} | retained-contract | missing | hidden-recipe-driven | undeclared | high | e2e-justification, unit-coverage-reviewed, end-state oracle |
 
-**Action:** Any TC with missing evidence, helper-artifact drift, hidden recipes, workaround dependence, or unsupported internal-detail checks should be updated during the next rewrite cycle.
+**Action:** Any TC with missing evidence, undeclared/wildcard artifact drift, hidden recipes, workaround dependence, or unsupported internal-detail checks should be updated during the next rewrite cycle.
 
 ### Gap Analysis
 

data/handbook/workflow-instructions/e2e/rewrite.wf.md

@@ -42,7 +42,9 @@ ace-bundle wfi://e2e/review  →  ace-bundle wfi://e2e/plan-changes  →  ace-bu
 - Normalize runner files to execution-only language.
 - Normalize verifier files to verdict-only, impact-first validation.
 - Keep setup concerns in `scenario.yml` and fixtures, not in TC runner setup sections.
-- Remove helper artifact requirements from `results/tc/{NN}/`; use runner observations instead.
+- Keep only declared verifier-dependent evidence under `results/tc/{NN}/`.
+- Move verifier-only artifact references into explicit runner/setup declarations.
+- Replace wildcard artifact paths with exact declared files.
 - Rewrite goal-style TCs around the public user path. Do not preserve hidden recipes, workaround branches, or supporting-tool probes as the way the runner reaches the goal.
 
 ## Workflow Steps
@@ -128,12 +130,14 @@ Follow the E2E test writing rules:
 - Target 2-5 TCs per scenario
 - Test through the CLI interface, not library imports
 - Write runner goals as “do the job” outcomes, not “write a report for the verifier” chores
-- Keep `results/tc/{NN}/` for real outcomes only; avoid helper YAML, path files, command files, and reflections
+- Keep `results/tc/{NN}/` for declared verifier-dependent evidence only; avoid undeclared helper YAML, reflections, and verifier-facing manifests
+- Keep only declared verifier-dependent evidence under `results/tc/{NN}/`; small supporting captures are acceptable when they are explicit and necessary
 - Use runner observations as the only non-filesystem secondary evidence source
 - Make final sandbox state or real product output the primary oracle whenever possible
 - Add behavioral/content assertions only when CLI output itself is part of the user-visible outcome
 - Remove duplicate command-only TCs; fold related assertions into one TC where possible
 - Do not encode exact workaround procedures, hidden command recipes, or internal debugging tricks the user would not infer from docs/usage/`--help`
+- For watch/live-output flows, rewrite to a bounded-session pattern with explicit shutdown evidence
 - If the job is valid but the public surface is too weak, plan a product/docs/help fix instead of hardcoding the workaround into the TC
 
 **Load the TC template for reference:**
@@ -152,6 +156,8 @@ For each TC classified as MODIFY:
    - **Broaden scope** — add assertions for related behavior tested by the same CLI invocation
    - **Fix structure** — add missing sections, fix formatting issues
    - **Replace helper-artifact oracles** — if the existing TC relies on runner-written helper files, rewrite it around final sandbox state plus runner observations
+   - **Declare verifier-dependent artifacts** — if the verifier names a `results/tc/...` file, ensure the runner or setup declares the exact same path
+   - **Remove wildcard declarations** — replace `results/tc/.../*` or `results/tc/.../foo.*` with exact paths
    - **Add evidence gates** — if the existing TC relies on existence-only or missing end-state checks, strengthen the primary oracle before falling back to debug captures
    - **Remove hidden recipes/workarounds** — if the existing TC teaches the runner how to bypass the public surface, rewrite it around the supported user path or narrow/remove the TC
 3. Update the `last-verified` field if the TC was re-run during modification
@@ -241,7 +247,8 @@ Present the execution summary:
 - [ ] TC count matches plan: {yes/no}
 - [ ] No stale references: {yes/no}
 - [ ] All scenarios have 2-5 TCs: {yes/no}
-- [ ] Modified/created TCs avoid helper files in `results/tc/{NN}/`: {yes/no}
+- [ ] Modified/created TCs avoid undeclared helper files in `results/tc/{NN}/`: {yes/no}
+- [ ] Modified/created TCs declare every verifier-dependent artifact path: {yes/no}
 
 ### Next Steps
 

data/lib/ace/test/end_to_end_runner/atoms/artifact_contract_validator.rb

@@ -0,0 +1,138 @@
+# frozen_string_literal: true
+
+module Ace
+  module Test
+    module EndToEndRunner
+      module Atoms
+        # Validates that verifier-visible artifact paths are explicitly declared by
+        # runner instructions or scenario setup, and normalizes grouped capture
+        # shorthand such as `foo.stdout`, `.stderr`, `.exit`.
+        class ArtifactContractValidator
+          Reference = Struct.new(:path, :optional, :source, :line, keyword_init: true)
+
+          FULL_PATH_PATTERN = /
+            (?:`|"|')?
+            (results\/tc\/\d{2}\/[^\s`)"']+|results\/tc\/\d{2}\/)
+            (?:`|"|')?
+            (\s*\(optional\))?
+          /ix
+          SUFFIX_PATTERN = /,\s*(?:`|"|')?(\.[A-Za-z0-9*._-]+)(?:`|"|')?(\s*\(optional\))?/i
+          WILDCARD_PATTERN = /[*?\[]/.freeze
+
+          class << self
+            def extract(markdown, source:)
+              markdown.to_s.each_line.with_index(1).flat_map do |line, line_number|
+                extract_from_line(line, source: source, line_number: line_number)
+              end
+            end
+
+            def references_from_paths(paths, source:)
+              Array(paths).filter_map do |path|
+                normalized = normalize(path)
+                next if normalized.nil?
+
+                Reference.new(path: normalized, optional: false, source: source, line: nil)
+              end
+            end
+
+            def validate!(tc_id:, scenario_dir:, runner_references:, verifier_references:, scenario_references:)
+              invalid_wildcards = (runner_references + verifier_references + scenario_references).select do |reference|
+                wildcard?(reference.path)
+              end
+              unless invalid_wildcards.empty?
+                raise ArgumentError,
+                  "Wildcard artifact path(s) are not supported for #{tc_id} in #{scenario_dir}: " \
+                  "#{format_references(invalid_wildcards)}"
+              end
+
+              declared_paths = normalized_paths(scenario_references + runner_references)
+              undeclared = verifier_references.reject do |reference|
+                declared_paths.include?(normalize(reference.path))
+              end
+              return if undeclared.empty?
+
+              raise ArgumentError,
+                "Verifier references undeclared artifact(s) for #{tc_id} in #{scenario_dir}: " \
+                "#{format_references(undeclared)}. " \
+                "Declare exact artifact paths in the runner file or scenario.yml sandbox-layout."
+            end
+
+            private
+
+            def extract_from_line(line, source:, line_number:)
+              matches = []
+              line.to_enum(:scan, FULL_PATH_PATTERN).each do
+                matches << {
+                  start: Regexp.last_match.begin(0),
+                  end: Regexp.last_match.end(0),
+                  path: normalize(Regexp.last_match[1]),
+                  optional: !Regexp.last_match[2].to_s.empty?
+                }
+              end
+
+              matches.each_with_index.flat_map do |match, index|
+                refs = [
+                  Reference.new(
+                    path: match[:path],
+                    optional: match[:optional],
+                    source: source,
+                    line: line_number
+                  )
+                ]
+
+                next_match = matches[index + 1]
+                suffix_region = line[match[:end]...(next_match ? next_match[:start] : line.length)].to_s
+                suffix_base = suffix_base_for(match[:path])
+                next refs if suffix_base.nil?
+
+                suffix_region.to_enum(:scan, SUFFIX_PATTERN).each do
+                  refs << Reference.new(
+                    path: "#{suffix_base}#{Regexp.last_match[1]}",
+                    optional: !Regexp.last_match[2].to_s.empty?,
+                    source: source,
+                    line: line_number
+                  )
+                end
+                refs
+              end
+            end
+
+            def suffix_base_for(path)
+              return nil if path.nil?
+              return nil if path.match?(%r{\Aresults/tc/\d{2}\z})
+
+              path.sub(/\.[^.\/]+\z/, "").tap do |value|
+                return nil if value == path
+              end
+            end
+
+            def normalized_paths(references)
+              references.map { |reference| normalize(reference.path) }.compact.uniq
+            end
+
+            def normalize(path)
+              value = path.to_s.strip
+              return nil unless value.start_with?("results/tc/")
+
+              value.sub(%r{/+\z}, "")
+            end
+
+            def wildcard?(path)
+              path.to_s.match?(WILDCARD_PATTERN)
+            end
+
+            def format_references(references)
+              references.uniq { |reference| [reference.path, reference.source, reference.line] }.map do |reference|
+                if reference.line
+                  "#{reference.path} (#{reference.source}:#{reference.line})"
+                else
+                  "#{reference.path} (#{reference.source})"
+                end
+              end.join(", ")
+            end
+          end
+        end
+      end
+    end
+  end
+end