githubnext/gh-aw-cao #11829 is still actionable: the optimization package page in target/optimization/dashboard.json does not match the issue’s requested read-only token-efficiency MVP, because it currently centers on candidate-readiness and intervention portfolio views instead of a 4-view operator dashboard for fleet AIC trend, highest-impact workflows, token diagnostics, and optimization value/churn. Confidence: high.
Action: A dashboard maintainer should either close this issue as superseded by a newer scoped replacement or assign it to Copilot to re-scope optimization/dashboard.json and its focused tests so the acceptance checks in #11829 pass; accept when the page shows the four requested evidence-backed views and preserves unavailable future evidence as unavailable rather than synthesized.
Evidence
target/optimization/dashboard.json defines the optimization-dashboard page with these essential views:
token-efficiency-portfolio-readiness (pie chart)
token-efficiency-portfolio-candidates (ranked portfolio candidate table)
token-efficiency-progress (pie chart)
- Those views focus on candidate readiness, ranked optimization candidates, and mature intervention progress, which differs materially from the issue’s requested MVP views:
- fleet AIC trend line chart,
- highest-impact workflows table,
- token and efficiency diagnostics split by token class,
- optimization value and churn table.
- The current dashboard file already contains canonical token-efficiency sources and query logic, including
token-efficiency-supported-run-smells, token-efficiency-opportunities, token-efficiency-interventions, and token-efficiency-comparisons, so this does not look blocked by total source absence; it looks like the page composition and focused coverage were never aligned to the issue.
target/tests/unit/optimization-token-optimizer-contract.test.mjs still asserts only that the dashboard contains token-efficiency-opportunities and token-efficiency-interventions sourced from audits; it does not verify the issue’s required page structure or the four-view MVP acceptance criteria.
- Issue dependency check:
#11800 is merged, so the stated dependency is satisfied.
- I found no materially equivalent open repo-assist triage issue already tracking
githubnext/gh-aw-cao issue 11829 triage guidance in the safe-output repository.
Rejected alternatives:
- Resolve as already fixed: not supported, because the visible page structure in
optimization/dashboard.json does not match the issue’s requested MVP.
- Clarify: not needed yet; the issue already contains concrete design, query, and acceptance requirements.
- Label-only action: weaker than the available evidence-backed investigation result.
Agent prompt
Update the optimization package dashboard to satisfy issue #11829 using a JSON-first Dashboard Language implementation. In optimization/dashboard.json, replace the current essential page composition with a phone-readable read-only token-efficiency dashboard containing exactly four essential views: (1) a fleet AIC trend line chart aggregated once per invocation from usage, (2) a highest-impact workflows ranked table grouped by canonical repository/workflow identity, (3) a token and efficiency diagnostics view that keeps input, output, cache-read, cache-write, and reasoning tokens separate and incorporates evidence-backed agent-smells, and (4) an optimization value and churn table that distinguishes proposed savings, gross realized savings, attributable overhead, net realized savings, verified net gain, disposition, and supersession lineage. Keep missing future producer evidence explicitly unavailable rather than synthesizing rows or zeros. Then update the focused unit tests, contract fixtures, and E2E coverage so every declared source resolves through the production worker/query boundary and the page covers loading, populated, empty, unavailable, and reactive-update states.
Control Plane
Generated by 🏷️ Repo Assist / Issue Triage · pi · gpt54 · 20.8 AIC · ⊞ 9.4K · ◷
githubnext/gh-aw-cao
#11829is still actionable: the optimization package page intarget/optimization/dashboard.jsondoes not match the issue’s requested read-only token-efficiency MVP, because it currently centers on candidate-readiness and intervention portfolio views instead of a 4-view operator dashboard for fleet AIC trend, highest-impact workflows, token diagnostics, and optimization value/churn. Confidence: high.Action: A dashboard maintainer should either close this issue as superseded by a newer scoped replacement or assign it to Copilot to re-scope
optimization/dashboard.jsonand its focused tests so the acceptance checks in#11829pass; accept when the page shows the four requested evidence-backed views and preserves unavailable future evidence as unavailable rather than synthesized.Evidence
target/optimization/dashboard.jsondefines theoptimization-dashboardpage with these essential views:token-efficiency-portfolio-readiness(pie chart)token-efficiency-portfolio-candidates(ranked portfolio candidate table)token-efficiency-progress(pie chart)token-efficiency-supported-run-smells,token-efficiency-opportunities,token-efficiency-interventions, andtoken-efficiency-comparisons, so this does not look blocked by total source absence; it looks like the page composition and focused coverage were never aligned to the issue.target/tests/unit/optimization-token-optimizer-contract.test.mjsstill asserts only that the dashboard containstoken-efficiency-opportunitiesandtoken-efficiency-interventionssourced fromaudits; it does not verify the issue’s required page structure or the four-view MVP acceptance criteria.#11800is merged, so the stated dependency is satisfied.githubnext/gh-aw-cao issue 11829 triage guidancein the safe-output repository.Rejected alternatives:
optimization/dashboard.jsondoes not match the issue’s requested MVP.Agent prompt
Update the optimization package dashboard to satisfy issue
#11829using a JSON-first Dashboard Language implementation. Inoptimization/dashboard.json, replace the current essential page composition with a phone-readable read-only token-efficiency dashboard containing exactly four essential views: (1) a fleet AIC trend line chart aggregated once per invocation fromusage, (2) a highest-impact workflows ranked table grouped by canonical repository/workflow identity, (3) a token and efficiency diagnostics view that keeps input, output, cache-read, cache-write, and reasoning tokens separate and incorporates evidence-backedagent-smells, and (4) an optimization value and churn table that distinguishes proposed savings, gross realized savings, attributable overhead, net realized savings, verified net gain, disposition, and supersession lineage. Keep missing future producer evidence explicitly unavailable rather than synthesizing rows or zeros. Then update the focused unit tests, contract fixtures, and E2E coverage so every declared source resolves through the production worker/query boundary and the page covers loading, populated, empty, unavailable, and reactive-update states.Control Plane
35430020435-4githubnext/gh-aw-cao