qualiow-exploratory-testing 2.0.0 → 2.2.0

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (43) hide show
  1. package/.claude-plugin/marketplace.json +26 -0
  2. package/.claude-plugin/plugin.json +1 -1
  3. package/CHANGELOG.md +65 -0
  4. package/README.md +123 -12
  5. package/agents/qa-diff-indexer-agent.md +81 -0
  6. package/agents/qa-gather-agent.md +21 -4
  7. package/agents/qa-page-mapper-agent.md +74 -0
  8. package/agents/qa-reporting-agent.md +103 -0
  9. package/bin/qualiow +30 -0
  10. package/data/knowledge/manifest.yml +22 -0
  11. package/dist/cli/index.js +876 -43
  12. package/dist/index.d.ts +50 -2
  13. package/dist/index.js +474 -29
  14. package/docs/GETTING-STARTED.md +127 -8
  15. package/hooks/hooks.json +27 -0
  16. package/hooks/scripts/read-guard.mjs +95 -0
  17. package/hooks/scripts/secret-patterns.d.mts +8 -0
  18. package/hooks/scripts/secret-patterns.mjs +136 -0
  19. package/hooks/scripts/write-guard.mjs +87 -0
  20. package/package.json +3 -2
  21. package/skills/qa-explore/SKILL.md +2 -1
  22. package/skills/qa-explore/phases/00-setup.md +16 -8
  23. package/skills/qa-explore/phases/03-discovery.md +48 -6
  24. package/skills/qa-explore/phases/07-reporting.md +36 -15
  25. package/skills/qa-explore/references/delegation-rules.md +82 -0
  26. package/skills/qa-explore/references/output-contract.md +7 -2
  27. package/skills/qa-explore/references/paths.md +25 -1
  28. package/skills/qa-explore/references/security-rules.md +4 -0
  29. package/skills/qa-explore/references/session-rules.md +2 -1
  30. package/skills/qa-explore-cleanup/SKILL.md +44 -19
  31. package/skills/qa-explore-feedback/SKILL.md +5 -0
  32. package/skills/qa-explore-mobile/SKILL.md +1 -1
  33. package/skills/qa-explore-mobile/phases/07-reporting.md +33 -11
  34. package/skills/qa-explore-quick/SKILL.md +12 -2
  35. package/skills/qa-explore-report/SKILL.md +29 -5
  36. package/skills/qa-gather/SKILL.md +30 -9
  37. package/skills/qa-knowledge-add/SKILL.md +10 -5
  38. package/skills/qa-knowledge-list/SKILL.md +28 -40
  39. package/skills/qa-verify-backend/SKILL.md +7 -4
  40. package/skills/qa-verify-backend/phases/02-static-review.md +11 -0
  41. package/skills/qa-verify-backend/phases/03-live-verification.md +3 -1
  42. package/skills/qa-verify-backend/phases/03b-api-verification.md +3 -1
  43. package/skills/qa-verify-backend/phases/05-reporting.md +51 -19
@@ -7,7 +7,9 @@ description: >
7
7
  Use when user says: "gather requirements", "prepare context", "analyze this ticket",
8
8
  "what should I test", or provides requirements documents before an explore session.
9
9
  argument-hint: "<file | url | text> [--domain <id>] [--output <file>]"
10
- allowed-tools: Read, Write, Glob, Grep, WebFetch, Bash(git:*)
10
+ allowed-tools: Read, Write, Glob, Grep, WebFetch, Bash(git:*), Bash(wc:*)
11
+ context: fork
12
+ agent: qa-gather-agent
11
13
  ---
12
14
 
13
15
  # Requirements Gathering & Analysis for Exploratory Testing
@@ -19,9 +21,11 @@ to `output/context/` (gitignored). Paths and the data-resolution order:
19
21
  `${CLAUDE_SKILL_DIR}/../qa-explore/references/paths.md`. The context file this skill writes
20
22
  is the input to a session whose artefacts follow
21
23
  `${CLAUDE_SKILL_DIR}/../qa-explore/references/output-contract.md` — the confidentiality
22
- header rule there applies to the context file too. For a long or multi-source gather, this
23
- skill can run in the background as the `qa-gather-agent` sub-agent
24
- (`.claude/agents/qa-gather-agent.md`), which follows this same process.
24
+ header rule there applies to the context file too. This skill runs in the
25
+ `qa-gather-agent` sub-agent (`qualiow:qa-gather-agent` under a plugin install), on Sonnet, in
26
+ a forked context: everything it needs must be in the invocation — file paths, URLs, or the
27
+ pasted text in the same message. There is no way to ask a follow-up question; every
28
+ ambiguity becomes a `[GAP]` / `[ASSUMPTION]` marker in the context file.
25
29
 
26
30
  You are a **Senior QA Analyst** preparing context for an exploratory testing session. Your job is to collect requirements from whatever sources are available, analyze them for completeness and testability, and produce a structured context file that the `/qa-explore` skill can consume.
27
31
 
@@ -58,7 +62,17 @@ You are a **Senior QA Analyst** preparing context for an exploratory testing ses
58
62
  ## Input Sources
59
63
 
60
64
  ### Source: Local File
61
- Read the file and extract:
65
+ Size it before you open it:
66
+
67
+ ```bash
68
+ wc -l <file>
69
+ ```
70
+
71
+ At most 300 lines — read it whole. Over 300 lines — `Grep '^#'` for the headings and `Read`
72
+ only the sections that carry requirements (description, acceptance criteria, out-of-scope,
73
+ open questions), with `offset`/`limit`. Never read a long specification end to end.
74
+
75
+ Extract:
62
76
  - Feature description
63
77
  - Acceptance criteria (look for "AC:", "Given/When/Then", numbered criteria, checkboxes)
64
78
  - Technical details (APIs, data models, components mentioned)
@@ -79,12 +93,15 @@ Parse the user's inline text for requirements information. People paste:
79
93
  - Verbal descriptions of what the feature should do
80
94
 
81
95
  ### Source: Git/PR
82
- If a git repo is available:
96
+ If a git repo is available, take the shape of the change first and the content second:
83
97
  ```bash
84
- git log --oneline -20 # Recent changes
85
- git diff main...HEAD # What changed on this branch
98
+ git log --oneline -20 # Recent changes
99
+ git diff --stat main...HEAD # What changed on this branch — file names and sizes only
100
+ git diff main...HEAD -- <path> # One file that maps to the ticket
86
101
  ```
87
- Extract: files changed, components affected, commit messages describing intent.
102
+ Never diff the whole branch into context. Read the `--stat` output, pick the files that map
103
+ to the ticket, and diff those one at a time. Extract: files changed, components affected,
104
+ commit messages describing intent.
88
105
 
89
106
  ### Source: Multiple Combined
90
107
  When multiple sources are provided, merge them into a unified context. Resolve conflicts by noting both versions.
@@ -116,9 +133,13 @@ Compare extracted requirements against the relevant domain checklist (read from
116
133
  [GAP] No mention of mobile/responsive behavior
117
134
  [GAP] Acceptance criteria says "user can checkout" but doesn't specify: with what payment methods? Is address required?
118
135
  [AMBIGUITY] "Fast checkout" — what is "fast"? Under 3 seconds? Under 5?
136
+ [ASSUMPTION] Read "supported cards" as Visa and Mastercard only — the ticket never says
119
137
  [MISSING AC] No acceptance criteria for: empty cart, duplicate submission, session timeout
120
138
  ```
121
139
 
140
+ You cannot ask the user a question from a forked context: anything you had to decide to keep
141
+ going is an `[ASSUMPTION]` line, named so the session can challenge it.
142
+
122
143
  ### Step 3: Extract Test Scenarios
123
144
  From the requirements, derive:
124
145
  - **Happy path scenarios** (what should work)
@@ -6,7 +6,7 @@ description: >
6
6
  Use when user says: "add knowledge", "add heuristic", "learn this", "save this technique",
7
7
  "add to knowledge base", or provides QA documentation to ingest.
8
8
  argument-hint: "[<text> | <url> | <file>]"
9
- allowed-tools: Read, Write, Glob, Grep, WebFetch, Bash(node:*), Bash(npx qualiow:*), Bash(qualiow:*)
9
+ allowed-tools: Read, Write, Glob, Grep, WebFetch, Bash(node:*), Bash(npx qualiow:*), Bash(npx:*), Bash(qualiow:*)
10
10
  ---
11
11
 
12
12
  # Add QA Knowledge
@@ -76,7 +76,8 @@ Present the structured entry and ask:
76
76
 
77
77
  ### 5. Determine Release Version
78
78
 
79
- Read `<data>/knowledge/manifest.yml` for current version.
79
+ `Grep '^version:' <data>/knowledge/manifest.yml` for the current version — never read the
80
+ manifest whole.
80
81
  - If adding to existing release: place in current version's entries/
81
82
  - If user wants a new release: bump version, create new release directory
82
83
 
@@ -87,12 +88,16 @@ Default: add to current release unless user specifies otherwise.
87
88
  1. Write the YAML entry:
88
89
  - inside this repository: `data/knowledge/releases/<version>/entries/<id>.yml`, and add its id to that release's `release.yml` `entries:` list (bump `entry_count`)
89
90
  - in a consumer project: `$PWD/data/knowledge/custom/<id>.yml`
90
- 2. Update `data/knowledge/manifest.yml`: run `npx qualiow kb sync` (regenerates the `entries:`
91
+ 2. Update `data/knowledge/manifest.yml`: run `qualiow kb sync` (regenerates the `entries:`
91
92
  registry and `stats` from the release files); if the CLI is unavailable, add the entry to
92
93
  `entries:` by hand with `id`, `file`, `type`, `priority`, `tags`, `domains` and bump `stats`.
93
- Then add the id to `loading_strategy` (`always`, `by_domain`, or `by_tag`).
94
+ Then add the id to `loading_strategy` (`always`, `by_domain`, `by_tag`, or `by_skill`).
94
95
  3. Update `data/knowledge/changelog.yml`: add the entry to the current release's `entries_added` list.
95
- 4. Verify: `npx qualiow validate --all` (or `npx qualiow kb check`) must pass.
96
+ 4. Verify: `qualiow validate --all` (or `qualiow kb check`) must pass.
97
+
98
+ Resolve the `qualiow` prefix per `${CLAUDE_SKILL_DIR}/../qa-explore/references/paths.md` and
99
+ write it literally; in a project that depends on the npm package that is
100
+ `npx -y -p qualiow-exploratory-testing qualiow`, never bare `npx qualiow`.
96
101
 
97
102
  ### 7. Confirm
98
103
 
@@ -6,58 +6,46 @@ description: >
6
6
  Use when user says: "list knowledge", "show heuristics", "what's in the knowledge base",
7
7
  "knowledge stats", "show changelog".
8
8
  argument-hint: "[--domain <id>] [--tag <tag>] [--type <type>] [--entry <id>] [--changelog] [--stats]"
9
- allowed-tools: Read, Glob, Grep
9
+ allowed-tools: Read, Grep, Bash(qualiow:*), Bash(npx:*)
10
10
  ---
11
11
 
12
12
  # Browse QA Knowledge Base
13
13
 
14
- Read the base from `<data>/knowledge/` (resolution order in
15
- `${CLAUDE_SKILL_DIR}/../qa-explore/references/paths.md`), plus any `$PWD/data/knowledge/custom/*.yml`.
14
+ This skill is a thin wrapper around the `qualiow` CLI, which reads the base from
15
+ `<data>/knowledge/` plus any `$PWD/data/knowledge/custom/*.yml`. Resolve the `qualiow` prefix
16
+ once per `${CLAUDE_SKILL_DIR}/../qa-explore/references/paths.md` and write it literally.
17
+
16
18
  Never paste a full entry into a session report or a website; the knowledge base is internal
17
19
  (`${CLAUDE_SKILL_DIR}/../qa-explore/references/security-rules.md`). Cite an entry by id in
18
20
  the `## Observations` section of a session report instead — see
19
21
  `${CLAUDE_SKILL_DIR}/../qa-explore/references/output-contract.md`.
20
22
 
21
- ## Commands
23
+ ## Flag mapping
22
24
 
23
- ### List All
24
- Read `<data>/knowledge/manifest.yml` and display:
25
- - Total entries by type (heuristics, techniques, checklists, references, patterns), counted
26
- from the `entries:` registry; if `stats` disagrees with the count, say so (run `npx qualiow kb sync`)
27
- - Active releases
28
- - Last updated date
25
+ | The user asks for | Run |
26
+ |---|---|
27
+ | everything | `qualiow list knowledge` |
28
+ | `--domain <id>` | `qualiow list knowledge --domain <id>` |
29
+ | `--tag <tag>` | `qualiow list knowledge --tag <tag>` |
30
+ | `--type <type>` | `qualiow list knowledge --type <type>` |
31
+ | `--entry <id>` | `qualiow list knowledge --entry <id>` |
32
+ | `--changelog` | `qualiow list knowledge --changelog` |
33
+ | `--stats` | `qualiow list knowledge --stats` |
29
34
 
30
- Then list all entries grouped by type:
31
- ```
32
- ## Heuristics (N)
33
- - SFDIPOT - Product Elements (v0.1.0) [high] [tags: exploration, what-to-test]
34
- - FEW HICCUPPS - Consistency Oracles (v0.1.0) [high] [tags: oracles, recognize-problems]
35
- ...
35
+ Flags combine; pass them through exactly as given. `--type` ∈ `heuristic` · `technique` ·
36
+ `checklist` · `reference` · `pattern`.
36
37
 
37
- ## Techniques (N)
38
- ...
38
+ ```bash
39
+ qualiow list knowledge --domain fintech --tag security
40
+ qualiow list knowledge --entry heuristic-sfdipot
39
41
  ```
40
42
 
41
- ### Filter by Domain
42
- `/qa-knowledge-list --domain ecommerce`
43
- Show only entries where `domains` includes the specified domain or "all".
44
-
45
- ### Filter by Tag
46
- `/qa-knowledge-list --tag security`
47
- Show only entries with the specified tag.
48
-
49
- ### Filter by Type
50
- `/qa-knowledge-list --type heuristic`
51
- Show only entries of the specified type.
52
-
53
- ### Show Entry Detail
54
- `/qa-knowledge-list --entry heuristic-sfdipot`
55
- Read and display the full YAML entry content.
56
-
57
- ### Show Changelog
58
- `/qa-knowledge-list --changelog`
59
- Read `<data>/knowledge/changelog.yml` and display recent changes.
43
+ ## Rules
60
44
 
61
- ### Show Stats
62
- `/qa-knowledge-list --stats`
63
- Display from manifest: entry counts, release history, loading strategy summary.
45
+ - Print the command output verbatim. Do not re-order it, re-summarize it, or add entries
46
+ from memory.
47
+ - Never `Read` `<data>/knowledge/manifest.yml` or a release entry directly — `--entry <id>`
48
+ returns one entry in full, and `Grep` answers a single-line question. The reasoning behind
49
+ this is in `${CLAUDE_SKILL_DIR}/../qa-explore/references/delegation-rules.md`.
50
+ - If the CLI is unavailable, say so and offer to `Grep` the manifest for what was asked
51
+ rather than reading it whole.
@@ -11,7 +11,7 @@ description: >
11
11
  "test the API", "check the endpoint", or gives a backend/API/infra ticket whose ACs
12
12
  cannot be seen in a browser.
13
13
  argument-hint: "[--target <id>] [--context <file>] [--static-only] [--api-only] [--no-e2e] [--parity <target-id>]"
14
- allowed-tools: Read, Write, Glob, Grep, Bash(git:*), Bash(aws:*), Bash(jq:*), Bash(grep:*), Bash(curl:*), Bash(node:*), Bash(playwright-cli:*), Bash(npx playwright-cli:*), Bash(terraform validate:*), Bash(terraform fmt:*)
14
+ allowed-tools: Read, Write, Glob, Grep, Bash(git:*), Bash(aws:*), Bash(jq:*), Bash(grep:*), Bash(curl:*), Bash(node:*), Bash(playwright-cli:*), Bash(npx playwright-cli:*), Bash(qualiow:*), Bash(npx:*), Bash(terraform validate:*), Bash(terraform fmt:*)
15
15
  ---
16
16
 
17
17
  # Backend & Infrastructure AC Verification
@@ -93,9 +93,12 @@ Execute in order. Read and follow the linked file.
93
93
 
94
94
  ## Knowledge Base
95
95
 
96
- Load these from `<data>/knowledge/` (resolved per `paths.md`) before the static and live lanes. They are the
97
- BE/API verification layer — techniques derived from real sessions with this skill,
98
- not from published literature.
96
+ Before the static and live lanes, run `qualiow kb digest --for backend` (resolve the prefix
97
+ per `${CLAUDE_SKILL_DIR}/../qa-explore/references/paths.md`). It prints a summary of each
98
+ entry below; fetch one in full with `qualiow list knowledge --entry <id>` when you need its
99
+ steps, and never `Read` the manifest or a release entry whole. The table is the reference
100
+ list of what the digest loads — the BE/API verification layer, techniques derived from real
101
+ sessions with this skill, not from published literature.
99
102
 
100
103
  | Entry | Use it when |
101
104
  |-------|-------------|
@@ -13,6 +13,17 @@ Group changed files by AC. Any file group that maps to **no** AC is either scope
13
13
  or an undocumented dependency — both worth a line in the report. Any AC that maps to
14
14
  **no** file is a candidate `FAIL` before you have read a single line.
15
15
 
16
+ **Gate on the size of the change.** 25 or more files, or 1,500 or more changed lines, in
17
+ the `--stat` output: invoke the `qa-diff-indexer-agent` sub-agent
18
+ (`qualiow:qa-diff-indexer-agent` under a plugin install) with the repo path, the base, the
19
+ branch and the AC list from phase 1. It returns an index — file, symbols or resources
20
+ touched, line ranges, candidate AC ids, note — plus the two lists that matter here: files
21
+ matching no AC, and ACs matching no file. It returns no verdicts and no severities; the
22
+ grouping above is still yours. Then read **only the ranges it names**, with
23
+ `git show <branch>:<path>`. Never diff the whole branch into context. If sub-agents are
24
+ unavailable, do the step yourself as in 2.1.0: read the diff by `--stat` first, then per
25
+ file.
26
+
16
27
  ## Step 2: Read the Implementation Against Each Claim
17
28
 
18
29
  Work AC by AC. For each, open the actual resource or handler and compare it to the
@@ -3,7 +3,9 @@
3
3
  Skip if the live lane is BLOCKED — but still write every probe command into
4
4
  `evidence/live-probes.md` so a human with credentials can run them unchanged.
5
5
 
6
- Read `${CLAUDE_SKILL_DIR}/references/aws-readonly-probes.md` for the command catalogue. Read
6
+ `${CLAUDE_SKILL_DIR}/references/aws-readonly-probes.md` is the command catalogue: Grep it for
7
+ the heading of the resource you need (`Grep '^## ' <file>` lists them) and Read only that
8
+ section with `offset`/`limit`; never the whole file. Read
7
9
  `${CLAUDE_SKILL_DIR}/references/safety-rules.md` before the first probe, and confirm the fingerprint from
8
10
  phase 0 step 3c is on file — a probe of an environment that does not run the change
9
11
  measures something else.
@@ -8,7 +8,9 @@ Run it whenever the ticket touches a request/response contract: a search or list
8
8
  endpoint, validation rules, pagination, filter values, an error contract, a payload
9
9
  shape a consumer reads. Skip it when the change has no HTTP surface at all.
10
10
 
11
- Read `${CLAUDE_SKILL_DIR}/references/api-probes.md` for the mechanics and the case catalogue,
11
+ `${CLAUDE_SKILL_DIR}/references/api-probes.md` holds the mechanics and the case catalogue:
12
+ Grep it for the heading of the section you need (`Grep '^## ' <file>` lists them) and Read
13
+ only that section with `offset`/`limit`; never the whole file. Read
12
14
  `${CLAUDE_SKILL_DIR}/references/environment-fingerprinting.md` before you believe any result, and
13
15
  `${CLAUDE_SKILL_DIR}/references/payload-verification.md` once you have one — a well-shaped `200` says nothing
14
16
  about whether the numbers in it are right.
@@ -130,18 +130,26 @@ ranked by what real users can reach **today** rather than by how bad each one re
130
130
  Close it with the same content in plain English, ready to paste into a ticket comment. That
131
131
  section is what gets the work scheduled.
132
132
 
133
- ## Step 4: Session Report
134
-
135
- Write `session-report.md` in the format of
136
- `${CLAUDE_SKILL_DIR}/../qa-explore/references/output-contract.md` with `Kind: backend`, and
137
- insert `## AC Matrix Summary` right after the executive summary:
138
- `n PASS / n PARTIAL / n FAIL / n BLOCKED / n NOT-REACHABLE / n UNVERIFIABLE` — all six
139
- verdicts, every time, so that what could not be observed stays visible. Then the findings
140
- ranked by severity, then what you could not check and exactly what access would unblock it.
141
- The coverage map uses the contract's `| Area | Risk | Status | Bugs | Notes |` header with
142
- `code-verified-only` for anything whose only evidence is a code reading. Write `stats.json`
143
- with `"kind": "backend"` and the verdict counts under `coverage.verdicts`, then append the
144
- rows to `output/sessions/INDEX.md` and `output/bugs/all-bugs.md` (columns in `paths.md`).
133
+ ## Step 4: Session Notes
134
+
135
+ The matrix, the bugs, the spec and the DoD findings above are yours. So is the judgement that
136
+ goes into the report — but not the assembling of it. Write
137
+ `phase-7-notes.md` (the same file name the reporting agent reads in every session kind), the
138
+ confidentiality header first, then exactly these sections:
139
+
140
+ - `## Executive Summary` — 3 sentences: what was verified, what was found, the biggest risk
141
+ - `## AC Matrix Summary` — `n PASS / n PARTIAL / n FAIL / n BLOCKED / n NOT-REACHABLE /
142
+ n UNVERIFIABLE`, all six verdicts, every time, so that what could not be observed stays
143
+ visible; it goes into the report right after the executive summary
144
+ - `## Coverage Map` — rows for `| Area | Risk | Status | Bugs | Notes |`, with
145
+ `code-verified-only` for anything whose only evidence is a code reading
146
+ - `## Observations` — the findings ranked by severity
147
+ - `## Areas Not Tested` — what you could not check and exactly what access would unblock it
148
+ - `## Recommendations` — closing with the disposition below
149
+ - `## Reflection`
150
+
151
+ Write `stats.json` with `"kind": "backend"` and the verdict counts under `coverage.verdicts`;
152
+ the report and the index rows are written in Step 5.
145
153
 
146
154
  Recommend a disposition in one line: ship, ship-with-follow-ups, or hold — and say what
147
155
  would change your mind. Where the scope of what you checked is narrower than the question
@@ -153,11 +161,35 @@ For a release-level session covering many tickets against one build, follow
153
161
  table first, then a fixed result vocabulary per ticket that keeps *not testable here* and
154
162
  *not tested* visible rather than letting them vanish between passed and failed.
155
163
 
156
- ## Step 5: Confidentiality
164
+ ## Step 5: Assemble, Finalize and Confidentiality
157
165
 
158
- Every output file starts with the confidentiality header from `output-contract.md`. Scan the
159
- whole session directory against the redaction list in `security-rules.md` (tokens, keys,
160
- cookies, real user emails, consumer PII, account ids not already in the target) before
161
- finishing; replace with `[REDACTED]`. Session output stays local; never transmit it anywhere.
162
- If a browser session was opened for the API lane, end it now: `playwright-cli -s=<sid> close`
163
- then `playwright-cli -s=<sid> delete-data`.
166
+ Invoke the `qa-reporting-agent` sub-agent (`qualiow:qa-reporting-agent` under a plugin
167
+ install) with the session directory and `kind: backend`. It assembles `session-report.md` from
168
+ `phase-7-notes.md`, `ac-matrix.md`, the bugs and `stats.json`, copying your summary,
169
+ recommendations and reflection verbatim, and then runs `qualiow session finalize`.
170
+
171
+ That scans every artefact with the same redaction list `src/utils/redact.ts` implements — the
172
+ list in `security-rules.md`: tokens, keys, cookies, real user emails, consumer PII, account
173
+ ids not already in the target — and the raw probe output under `evidence/` is scanned like
174
+ everything else. It also checks the confidentiality header on every markdown file and
175
+ validates `stats.json`, then appends the session row to `output/sessions/INDEX.md` (`Kind` =
176
+ `backend`) and one row per bug to `output/bugs/all-bugs.md`, and records the session metrics.
177
+ It is idempotent.
178
+
179
+ Read its return (at most 8 lines). If it reports a missing notes section or a violation it
180
+ could not fix, fix that yourself — replace the value with `[REDACTED]`, add the missing
181
+ header, fix the stats key — and re-run the check:
182
+
183
+ ```bash
184
+ qualiow session finalize output/sessions/<session-dir> --check
185
+ ```
186
+
187
+ `--check` validates and writes nothing, naming each offending file. Resolve the `qualiow`
188
+ prefix per `${CLAUDE_SKILL_DIR}/../qa-explore/references/paths.md`.
189
+
190
+ If sub-agents are unavailable, do the step yourself as in 2.1.0: write `session-report.md`
191
+ from the notes in the contract's format and run `qualiow session finalize output/sessions/<session-dir>`.
192
+
193
+ Session output stays local; never transmit it anywhere. If a browser session was opened for
194
+ the API lane, end it now: `playwright-cli -s=<sid> close` then
195
+ `playwright-cli -s=<sid> delete-data`.