gaia-submission

GAIA Submission Skill

Walk Claude Code through every step needed to go from a clean environment to a signed, HAL-compatible submission package ready to upload to the Princeton GAIA leaderboard.

When to use

When the user wants to:

• Run a benchmark and submit results to the HAL leaderboard
• Package an existing results file into a submission archive
• Confirm their environment is ready for a benchmark run

Prerequisites

Before starting, confirm these are available:

Requirement	Check
ANTHROPIC_API_KEY	echo ${ANTHROPIC_API_KEY:0:8}… (should show sk-ant-…)
HF_TOKEN	echo ${HF_TOKEN:0:5}… (should show hf_…)
Node.js 20+	node --version
CLI built	node v3/@swarmdo/cli/bin/cli.js --version

Phase 1 — Validate environment

# Run all pre-flight checks
/gaia validate

If any check fails, resolve it before continuing.

Phase 2 — Estimate cost and confirm

Ask the user for their configuration:

• Level (default: 1)
• Question limit (default: 53 for a quick run, 165 for the full L1 set)
• Models (default: claude-sonnet-4-6)
• Self-consistency voting (default: 1; use 3 for L2/L3)

/gaia cost --level=$LEVEL --limit=$LIMIT --models=$MODELS --voting=$VOTING

If projected cost > $5, show the estimate and ask: "This run will cost approximately $X. Proceed? (y/N)"

Phase 3 — Run the benchmark

/gaia run --level=$LEVEL --limit=$LIMIT --models=$MODELS --voting=$VOTING

While running, progress is reported every 5 questions:

[12/53] 22.7% (5 passed of 22 scored) — est. remaining: $0.18

Store the run summary in memory for history tracking:

npx @swarmdo/cli@latest memory store \
  --namespace gaia-runs \
  --key "run-$(date +%Y%m%d-%H%M)" \
  --value '{"level":$LEVEL,"model":"$MODEL","total":$TOTAL,"passed":$PASSED,"pass_rate":$RATE,"est_cost_usd":$COST}'

Phase 4 — Package for submission

/gaia submit --results=~/.cache/swarmdo/gaia/results-latest.json

This produces:

submission-<date>-<sha>/
├── results.jsonl        ← HAL-compatible, one JSON per line
├── trajectories.jsonl   ← full agent traces
├── metadata.json        ← harness info, model, tool catalogue
├── manifest.md.json     ← Ed25519-signed witness
└── README.md            ← human summary + leaderboard comparison

Phase 5 — Compare and report

/gaia leaderboard --level=$LEVEL
/gaia history

Interpret the gap between swarmdo's score and the leaderboard top-10. Identify the primary failure mode (tool gap, reasoning miss, extraction bug) using the /gaia-debugging skill if needed.

Phase 6 — Persist learnings

npx @swarmdo/cli@latest hooks post-task \
  --task-id "gaia-submission-$(date +%Y%m%d)" \
  --success true \
  --train-neural true

Store any discovered patterns:

npx @swarmdo/cli@latest memory store \
  --namespace gaia-patterns \
  --key "submission-notes-$(date +%Y%m%d)" \
  --value "Level $LEVEL, $MODEL: $NOTES"

Extensibility note

This skill is intentionally structured to be benchmark-agnostic. The phase headers (validate → estimate → run → package → compare → learn) apply to SWE-bench, WebArena, and HumanEval with only phase 3-4 details changing.