Review Quality Gates
Shared obligations below; load step-specific procedures only when needed.
Load map
| Step |
Read inside this skill |
| Any review |
this file: provenance, conventions, evidence, delivery |
| Shared-context preparation/dispatch |
references/context-sharing.md: artifacts, metrics, raw mode |
| Consolidation |
references/verification-panel.md: lenses, gating, cost guard, critic |
| Lens, critic or reviewer briefs |
references/panel-and-critic-prompts.md |
| code-review dispatch and map injection |
references/code-review-agents.md |
| code-review output and optional HTML |
references/code-review-output.md |
code-review fix/commit remediation |
references/code-review-fix-loop.md |
Lens 0 uses declared senior-review:premise-auditor. Lenses 1-3, critic and
inline briefs use the isolated-worker binding, senior-review:isolated-worker,
inheriting the session model. If the host provides no such
binding, declare independent verification unavailable instead of inventing one.
Shared-Context Provenance Rule
Evidence derived from a shared artifact cannot independently corroborate the claims contained in that same artifact. N reviewers agreeing on a premise they were all given is one observation, not N.
- Consuming a Codebase Map or interconnect claim does not verify it. Re-derive
its support from evidence before basing a finding on it; preserve prior exposure.
- Agreement on a shared premise is an echo, raising neither confidence nor
severity. Consolidation labels it accordingly.
- No metric rewards agreement with shared context. Map utilization measures
consumption, never quality.
Reviewer Pipeline Conventions
Discovery reviewers read references/reviewer-common.md for evidence, coverage,
knowledge accounting and compact delivery; roles own domain methods. Provider
reviewers receive these conventions through their briefs:
- Scope budget: complete assigned checks or name unchecked scope and cause.
Read count and absence of findings never prove irrelevance or completion.
- No-findings protocol: "examined X, Y, Z: no issues" is valid Phase 3 input.
- Cross-Reviewer Notes: route other dimensions' observations through
## Cross-Reviewer Notes in Phase 3.
- Context citation: cite the
01-context.md reference a finding relies on, if any.
Quality metrics: references/context-sharing.md, ### Metrics.
Evidence Classes
Label each finding's decisive evidence, strongest first:
reproduced: test/run fails on candidate and passes on base (reverse for a
fix); record command, scratch path and both outcomes.
tool-reported: verbatim evidence-pack tool result; see
references/candidate-grounding.md in senior-review:review-preparation.
measured: harness, simulation or real logs/metrics; state method. Keep
temporary harnesses outside the work tree, removing them under the Delivery Gate.
derived: inferred from code, not execution; default for unlabeled evidence.
- A
derived number alone cannot justify Critical severity: measure it or cap at High.
- Neither a derived number nor "bounded"/"low traffic" alone justifies accepting
a finding. Establish what the user or operator sees, and when.
"None (silent)" escalates consequential harm; it never mitigates it.
- Lens 2 checks rates/costs against actual deduplication, caps and phase transitions.
- Lens 1 may reproduce only when its brief grants a run-owned disposable workspace
and names the runner and scratch path. Write a minimal test there, run base and
candidate, record both outcomes; remove it unless adopted as regression protection.
Failure to reproduce is counterevidence. Without the grant, read and reason;
retain the reviewer's evidence class.
Delivery Gate
- Every reviewer delivers or declares: before consolidation, collect each
spawned reviewer's findings file or explicit no-findings report. Nudge an idle
reviewer once. Continued silence: salvage partial output as
[undelivered -- collected by orchestrator], marking the dimension degraded
or not delivered when empty, never clean. Missing required deliveries/checks
remain incomplete regardless of flags.
- The work tree is left as found: compare complete pre-review and final
workspace snapshots. New status entries do not prove ownership. Workers record
created files, original absence and final hashes in an explicit manifest;
preserve/report foreign or changed files. Keep probes in the owned run root.
Before removing an unchanged owned temporary artifact, preserve its conditions,
commands, outcomes, errors and limitations in the native report; verify remaining
evidence links resolve. Preserve unresolved evidence and resume data. Broad
cleanup or Git clean is forbidden. Return manifest and report; persistent
retention belongs to the caller's consolidate path, never status-based deletion.
1---2name: review-quality-gates3description: >4 Verification panel, completeness critic, pipeline conventions and shared-context provenance rules for multi-reviewer consolidation.5 TRIGGER WHEN: running a branch/team review through the host entry point or code-review quality gates, or consolidating findings from parallel reviewers.6metadata:7 source: https://skillmd.com/skills/acaprino/review-quality-gates8---910# Review Quality Gates1112Shared obligations below; load step-specific procedures only when needed.1314## Load map1516| Step | Read inside this skill |17|---|---|18| Any review | this file: provenance, conventions, evidence, delivery |19| Shared-context preparation/dispatch | `references/context-sharing.md`: artifacts, metrics, raw mode |20| Consolidation | `references/verification-panel.md`: lenses, gating, cost guard, critic |21| Lens, critic or reviewer briefs | `references/panel-and-critic-prompts.md` |22| code-review dispatch and map injection | `references/code-review-agents.md` |23| code-review output and optional HTML | `references/code-review-output.md` |24| code-review `fix`/`commit` remediation | `references/code-review-fix-loop.md` |2526Lens 0 uses declared `senior-review:premise-auditor`. Lenses 1-3, critic and27inline briefs use the isolated-worker binding, `senior-review:isolated-worker`,28inheriting the session model. If the host provides no such29binding, declare independent verification unavailable instead of inventing one.3031## Shared-Context Provenance Rule3233> **Evidence derived from a shared artifact cannot independently corroborate the claims contained in that same artifact. N reviewers agreeing on a premise they were all given is one observation, not N.**34351. Consuming a Codebase Map or interconnect claim does not verify it. Re-derive36 its support from evidence before basing a finding on it; preserve prior exposure.372. Agreement on a shared premise is an **echo**, raising neither confidence nor38 severity. Consolidation labels it accordingly.393. No metric rewards agreement with shared context. Map utilization measures40 consumption, never quality.4142## Reviewer Pipeline Conventions4344Discovery reviewers read `references/reviewer-common.md` for evidence, coverage,45knowledge accounting and compact delivery; roles own domain methods. Provider46reviewers receive these conventions through their briefs:4748- **Scope budget**: complete assigned checks or name unchecked scope and cause.49 Read count and absence of findings never prove irrelevance or completion.50- **No-findings protocol**: "examined X, Y, Z: no issues" is valid Phase 3 input.51- **Cross-Reviewer Notes**: route other dimensions' observations through52 `## Cross-Reviewer Notes` in Phase 3.53- **Context citation**: cite the `01-context.md` reference a finding relies on, if any.54 Quality metrics: `references/context-sharing.md`, `### Metrics`.5556## Evidence Classes5758Label each finding's decisive evidence, strongest first:5960- **`reproduced`**: test/run fails on candidate and passes on base (reverse for a61 fix); record command, scratch path and both outcomes.62- **`tool-reported`**: verbatim evidence-pack tool result; see63 `references/candidate-grounding.md` in `senior-review:review-preparation`.64- **`measured`**: harness, simulation or real logs/metrics; state method. Keep65 temporary harnesses outside the work tree, removing them under the Delivery Gate.66- **`derived`**: inferred from code, not execution; default for unlabeled evidence.67681. A `derived` number alone cannot justify **Critical** severity: measure it or cap at High.692. Neither a derived number nor "bounded"/"low traffic" alone justifies accepting70 a finding. Establish **what the user or operator sees, and when**.71 "None (silent)" escalates consequential harm; it never mitigates it.723. Lens 2 checks rates/costs against actual deduplication, caps and phase transitions.734. Lens 1 may reproduce only when its brief grants a run-owned disposable workspace74 and names the runner and scratch path. Write a minimal test there, run base and75 candidate, record both outcomes; remove it unless adopted as regression protection.76 Failure to reproduce is counterevidence. Without the grant, read and reason;77 retain the reviewer's evidence class.7879## Delivery Gate80811. **Every reviewer delivers or declares:** before consolidation, collect each82 spawned reviewer's findings file or explicit no-findings report. Nudge an idle83 reviewer once. Continued silence: salvage partial output as84 `[undelivered -- collected by orchestrator]`, marking the dimension **degraded**85 or **not delivered** when empty, never clean. Missing required deliveries/checks86 remain incomplete regardless of flags.872. **The work tree is left as found:** compare complete pre-review and final88 workspace snapshots. New status entries do not prove ownership. Workers record89 created files, original absence and final hashes in an explicit manifest;90 preserve/report foreign or changed files. Keep probes in the owned run root.91 Before removing an unchanged owned temporary artifact, preserve its conditions,92 commands, outcomes, errors and limitations in the native report; verify remaining93 evidence links resolve. Preserve unresolved evidence and resume data. Broad94 cleanup or Git clean is forbidden. Return manifest and report; persistent95 retention belongs to the caller's consolidate path, never status-based deletion.