Esat
Tri-model (Claude + Codex + Grok) world-class level-up analysis. Same engine as eskill-analyze, but the stress-test phase is a mandatory 3-SOTA-model panel — Claude critic + Codex + Grok independently review the item, then Claude synthesizes consensus / divergence / unique catches. Fixed trio product (not a configurable roster). Use for high-stakes analysis you want triple-checked across frontier models. Triggers: /esat, eskill analyze trio, eskill-analyze-trio, eskill-trio, trio analysis, three-model review.From its SKILL.md
npx -y skills add Renn-Labs/eskill-analyze --skill esatAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
One thing to look at
- 1 stars1 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
SKILL.md
3.6 KB, as published. Nobody here has run it
eSkill: Analyze Trio (esat) — 3-SOTA-Model Level-Up Analyzer
esat = eskill-analyze + a mandatory tri-model review panel. It runs the full world-class analysis engine, then has three frontier models — Claude (critic), Codex, Grok — independently stress-test the item and the draft in parallel. You (Claude) are the only trusted integrator: you synthesize their verdicts into one panel and adjust the analysis accordingly.
Use esat when the stakes justify a triple-model cross-check (architecture bets, strategy calls, build-vs-buy, anything you want frontier-model consensus on). For a lighter single-model pass, use /esa (eskill-analyze).
Product boundary: Tier 2 is a deliberately fixed trio. Configurable rosters and harness-aware multi-route selection live in Phase 9 of esat-fleet / esat-frontier, not here.
Engine — reuse eskill-analyze verbatim
Run the entire eskill-analyze protocol for Input Parameters, Triage, Delegation, Framework Selection, Steps 1–8, and Phase P Premortem (when triage activates it). Read and follow, in order:
${CLAUDE_SKILL_DIR}/../eskill-analyze/SKILL.md— Input Parameters, Scope check, Triage gate, Delegation Strategy, Framework Selection, Comparison Mode, Project-Level Impact, Post-Analysis Routing. Apply the two esat overrides below.${CLAUDE_SKILL_DIR}/../eskill-analyze/references/triage-guide.md— classification${CLAUDE_SKILL_DIR}/../eskill-analyze/references/mental-models.md— model selection${CLAUDE_SKILL_DIR}/../eskill-analyze/references/analysis-protocol.md— Phase 0 + Steps 1–8 + Phase P Premortem (when triage activates it). Skip that file's "Phase 9: Stress Test" — esat replaces Phase 9 with the trio panel.${CLAUDE_SKILL_DIR}/../eskill-analyze/references/world-class-signals.md
esat overrides (only two differences from eskill-analyze)
- Phase 9 is the Tri-Model Panel (this skill's
references/trio-panel.md), not the single Claude critic. - Phase 9 is mandatory, not discretionary. The trio panel is the entire point of esat — triage may not skip it. (Kill-switch
ESKILL_PEER=0degrades it to Claude-only; see below.)
Phase 9 — Tri-Model Panel
After completing Steps 1–8 and Phase P (when activated) and producing the draft analysis, read and execute ${CLAUDE_SKILL_DIR}/references/trio-panel.md. That file specifies the three concurrent reviewers, the untrusted-input synthesis rules, the gates/kill-switches, draft ownership for standalone vs caller-owned modes, and the output block.
Output
Use the eskill-analyze output template (${CLAUDE_SKILL_DIR}/../eskill-analyze/assets/output-template.md) plus the ### Tri-Model Panel block defined in references/trio-panel.md. Place the panel directly above ### Stress Test Notes. The triage header should read esat (trio) so it's clear which skill produced the analysis.
Save the analysis output to .omc/plans/esat-{date}-{focus}.md.
Post-Analysis Routing
Same as eskill-analyze (sprint/overnight routing on the confirmed action list). See its SKILL.md "Post-Analysis Routing" section.