harness/
SESSIONS · 8 tracked

Session intelligence

SKILL EFFECTIVENESS

Which skills carry sessions

1/8 sessions used = consider stronger trigger
SkillUsedSessionsCoverage
superpowers:brainstorming11
superpowers:writing-plans11
repo-conventions11
observability11
helm-charts11
authentication11
frontend11
database11
ci-cd11
self-review11
BEHAVIOR PATTERNS

Detected anti-patterns

Recurring failure modes Claude exhibits
finalizer-race-condition-pattern-inherited1
resync-filter-too-narrow1
RECOMMENDATIONS

Actions to take

superpowers:brainstormingsuperpowers:brainstorming skill is underused (1/8 sessions) — consider reinforcing in CLAUDE.md

Fix →

superpowers:writing-planssuperpowers:writing-plans skill is underused (1/8 sessions) — consider reinforcing in CLAUDE.md

Fix →

repo-conventionsrepo-conventions skill is underused (1/8 sessions) — consider reinforcing in CLAUDE.md

Fix →

observabilityobservability skill is underused (1/8 sessions) — consider reinforcing in CLAUDE.md

Fix →

helm-chartshelm-charts skill is underused (1/8 sessions) — consider reinforcing in CLAUDE.md

Fix →

authenticationauthentication skill is underused (1/8 sessions) — consider reinforcing in CLAUDE.md

Fix →

frontendfrontend skill is underused (1/8 sessions) — consider reinforcing in CLAUDE.md

Fix →

databasedatabase skill is underused (1/8 sessions) — consider reinforcing in CLAUDE.md

Fix →

ci-cdci-cd skill is underused (1/8 sessions) — consider reinforcing in CLAUDE.md

Fix →

self-reviewself-review skill is underused (1/8 sessions) — consider reinforcing in CLAUDE.md

Fix →
HISTORY

Session log

SessionModelSummaryEventsSkillsViol.Date
#18claude-opus-4-74104/25/2026
#14claude-opus-4-6Added natural language search with Ollama/Mistral, keyword filtering, and attempted Zillow PX bypass. NL search fully working end-to-end. Zillow disabled due to PerimeterX blocking all automated access.9304/11/2026
#17claude-opus-4-6Designed, built, deployed, and migrated onto a new HomelabSecret CRD + standalone operator. End-to-end: brainstorm -> design spec -> implementation plan -> 14-task subagent-driven execution -> CI + ArgoCD deployment -> migration of 6 existing InfisicalSecret manifests -> debug/fix 2 latent bugs inherited from the authentik-operator template -> all 6 CRs reach Ready, ~90 lines of boilerplate removed, no service disruption.13024/11/2026
#12claude-opus-4-6E2E test session1004/10/2026
#11claude-opus-4-6Added GitHub Actions observability: deployed cpanato/github_actions_exporter via ArgoCD, added OTel scrape targets, created Grafana dashboard. Required three fix commits for service routing, metric naming, and scrape target discovery.11304/10/2026
#9claude-opus-4-6All remaining feature gaps completed, verified with typecheck, committed and pushed. CI/CD will auto-deploy.1004/10/2026
#8claude-opus-4-6Added regression detection MCP tool and enhanced dashboard to show regressing evals.1004/10/2026
#10claude-opus-4-6Full session: brainstormed, designed, implemented, and deployed the Run Telemetry Pipeline feature, then completed all remaining feature gaps across the agent-harness platform. 4 commits, 2 migrations, 5 new MCP tools, 1 new dashboard route, 7 gap fixes.11304/9/2026