Skip to content

amazon-science/ammo / ai_cli_session/.codex/skills/ammo

Profile and optimize GPU kernels for vLLM inference on NVIDIA GPUs. Use when targeting a specific model, hardware, dtype, parallelism, and workload to improve latency.

FileKindSize
RUNTIME.mdmarkdown4.9 KB
SKILL.mdskill md18.5 KB
agents/ammo-auditor.mdmarkdown9.1 KB
agents/ammo-champion.mdmarkdown9.6 KB
agents/ammo-delegate.mdmarkdown3.2 KB
agents/ammo-implementer.mdmarkdown13.8 KB
agents/ammo-investigator.mdmarkdown4.0 KB
agents/ammo-report.mdmarkdown2.9 KB
agents/ammo-researcher.mdmarkdown10.3 KB
agents/ammo-resolver.mdmarkdown4.4 KB
agents/ammo-transcript-monitor.mdmarkdown7.8 KB
agents/openai.yamlexecutablescript313 B
data/ATTRIBUTION.mdmarkdown1.4 KB
data/gsm8k_full.jsonexecutablescript735.5 KB
data/gsm8k_subset.jsonexecutablescript116.8 KB
eval/EVAL-SKILL.mdexecutablescript21.8 KB
eval/agents/transcript_grader.mdexecutablescript18.0 KB
eval/causal/__init__.pyexecutablescript137 B
eval/causal/agents/causal_analyzer.mdexecutablescript10.5 KB
eval/causal/build_dag.pyexecutablescript21.6 KB
eval/causal/diff_versions.pyexecutablescript38.5 KB
eval/causal/extract_events.pyexecutablescript36.3 KB
eval/causal/generate_postmortem.pyexecutablescript17.0 KB
eval/causal/score_nodes.pyexecutablescript21.5 KB
eval/causal/templates/narrative_template.mdexecutablescript472 B
eval/causal/templates/viz_template.htmlexecutablescript9.5 KB
eval/causal/tests/__init__.pyexecutablescript151 B
eval/causal/tests/conftest.pyexecutablescript448 B
eval/causal/tests/fixtures/sample_session.jsonlexecutablescript10.4 KB
eval/causal/tests/fixtures/sample_session_data.jsonexecutablescript3.4 KB
eval/causal/tests/fixtures/sample_snapshot.jsonexecutablescript2.5 KB
eval/causal/tests/test_build_dag.pyexecutablescript2.7 KB
eval/causal/tests/test_diff_versions.pyexecutablescript13.5 KB
eval/causal/tests/test_extract_events.pyexecutablescript3.7 KB
eval/causal/tests/test_generate_postmortem.pyexecutablescript3.1 KB
eval/causal/tests/test_score_nodes.pyexecutablescript3.3 KB
eval/scripts/aggregate_versions.pyexecutablescript15.2 KB
eval/scripts/archive_run.pyexecutablescript9.1 KB
eval/scripts/codex_session.pyexecutablescript10.6 KB
eval/scripts/generate_dashboard.pyexecutablescript5.3 KB
eval/scripts/parse_artifacts.pyexecutablescript48.1 KB
eval/scripts/parse_session_logs.pyexecutablescript51.3 KB
eval/scripts/rescore_archived.pyexecutablescript5.5 KB
eval/scripts/score_campaign.pyexecutablescript43.8 KB
eval/scripts/snapshot_changes.pyexecutablescript11.1 KB
eval/scripts/verify_citations.pyexecutablescript8.8 KB
eval/tests/test_dashboard_bugs.pyexecutablescript8.8 KB
eval/tests/test_parse_session_logs.pyexecutablescript3.9 KB
eval/tests/test_parse_session_logs_round_inference.pyexecutablescript11.0 KB
eval/tests/test_score_campaign_diluted.pyexecutablescript9.4 KB
eval/viewer/dashboard_template.htmlexecutablescript15.1 KB
instrumentation/sitecustomize.pyexecutablescript3.6 KB
orchestration/audit-protocol.mdmarkdown9.3 KB
orchestration/debate-protocol.mdmarkdown11.2 KB
orchestration/integration-logic.mdmarkdown10.3 KB
orchestration/parallel-tracks.mdmarkdown8.9 KB
references/artifact-layout.mdexecutablescript22.4 KB
references/audit-invariants.mdmarkdown10.6 KB
references/champion-common-patterns.mdmarkdown3.4 KB
references/crossover-probing.mdexecutablescript4.7 KB
references/cudagraph-safety.mdmarkdown2.7 KB
references/debate-rules.mdmarkdown4.7 KB
references/debate-scoring-rubric.mdexecutablescript16.7 KB
references/e2e-delta-math.mdexecutablescript17.9 KB
references/e2e-latency-guide.mdexecutablescript9.6 KB
references/fusion-feasibility-heuristics.mdmarkdown5.6 KB
references/gpu-pool.mdexecutablescript5.4 KB
references/impl-track-rules.mdexecutablescript13.8 KB
references/nsys-profiling-guide.mdmarkdown9.8 KB
references/optimization-categories.mdexecutablescript11.0 KB
references/shutdown-protocol.mdexecutablescript3.1 KB
references/technology-selection.mdmarkdown5.0 KB
references/torch-compile-contract.mdmarkdown13.8 KB
references/validation-defaults.mdexecutablescript34.2 KB
references/workload-invocation.mdexecutablescript4.9 KB
references/writing-style.mdexecutablescript6.9 KB
scripts/ammo_state.pyexecutablescript142.0 KB
scripts/artifact_layout.jsonconfig16.3 KB
scripts/check_projection_accuracy.pyexecutablescript11.7 KB
scripts/create_evidence_template.pyexecutablescript9.7 KB
scripts/create_worktree_with_build.shexecutablescript10.2 KB
scripts/generate_validation_report.pyexecutablescript35.7 KB
scripts/gpu_reservation.pyexecutablescript27.1 KB
scripts/hook_cmd_classify.pyscript13.3 KB
scripts/mine_trace.pyscript29.4 KB
scripts/monitor_queue_ack.pyexecutablescript6.1 KB
scripts/ncu_sanity.shexecutablescript5.2 KB
scripts/ncu_sanity_driver.pyexecutablescript8.5 KB
scripts/new_target.pyexecutablescript22.0 KB
scripts/preflight_vllm_api.pyexecutablescript12.3 KB
scripts/reconcile_track_state.pyexecutablescript15.8 KB
scripts/remove_worktree_cleanup.shexecutablescript1.1 KB
scripts/resolve_codex_transcript.pyexecutablescript3.6 KB
scripts/run_vllm_bench_latency_sweep.pyexecutablescript249.6 KB
scripts/transcript_filter.pyexecutablescript38.4 KB
scripts/transitions.jsonexecutablescript1.5 KB
scripts/verify_validation_gates.pyexecutablescript88.8 KB
tests/agents/test-champion.mdexecutablescript20.4 KB
tests/agents/test-impl-champion.mdexecutablescript14.3 KB
tests/agents/test-implementer.mdexecutablescript15.8 KB
tests/agents/test-orchestrator.mdexecutablescript33.8 KB
tests/agents/test-researcher.mdexecutablescript21.0 KB
tests/agents/test-transcript-monitor.mdexecutablescript13.4 KB
tests/fixtures/diluted_pass/legacy_no_phase.golden.mdmarkdown1.1 KB
tests/fixtures/diluted_pass/legacy_no_phase.golden.summary.jsonconfig582 B
tests/fixtures/diluted_pass/legacy_no_phase.jsonconfig400 B
tests/fixtures/diluted_pass/legacy_scalar_phase.golden.mdmarkdown2.0 KB
tests/fixtures/diluted_pass/legacy_scalar_phase.golden.summary.jsonconfig1.0 KB
tests/fixtures/diluted_pass/legacy_scalar_phase.jsonconfig626 B
tests/fixtures/mining/real_campaign_bottleneck_analysis.mdmarkdown20.7 KB
tests/run_all.shexecutablescript1.6 KB
tests/scripts/test_transcript_filter.pyexecutablescript41.1 KB
tests/test_agent_runtime_contracts.pyexecutablescript1.8 KB
tests/test_ammo_state.pyexecutablescript71.7 KB
tests/test_artifact_layout_manifest.pyscript11.6 KB
tests/test_audit_started.pyscript18.2 KB
tests/test_correctness_comparator.pyscript12.6 KB
tests/test_diluted_pass.pyscript24.4 KB
tests/test_drift_compat.pyscript5.1 KB
tests/test_evidence_freshness.pyexecutablescript5.6 KB
tests/test_fail_closed.pyexecutablescript14.7 KB
tests/test_gpu_reservation.pyexecutablescript24.8 KB
tests/test_gpu_reservation_session_id.pyexecutablescript11.2 KB
tests/test_guidance_semantic_contract.pyscript24.9 KB
tests/test_hook_cmd_classify_parity.pyscript11.7 KB
tests/test_hook_semantics.pyexecutablescript95.3 KB
tests/test_mine_trace.pyscript24.2 KB
tests/test_mining_invalidated_schema.pyexecutablescript9.4 KB
tests/test_new_target_dp_ep.pyexecutablescript7.4 KB
tests/test_new_target_v2_scaffold.pyexecutablescript13.1 KB
tests/test_parse_artifacts.pyexecutablescript3.0 KB
tests/test_policy_convergence.pyscript797 B
tests/test_preflight_vllm_api.pyscript13.3 KB
tests/test_projection_accuracy_v2.pyexecutablescript4.1 KB
tests/test_reconcile_track_state.pyexecutablescript18.2 KB
tests/test_resolve_codex_transcript.pyexecutablescript3.1 KB
tests/test_state_schema_semantics.pyexecutablescript7.6 KB
tests/test_status_set_sync.pyexecutablescript9.1 KB
tests/test_sweep_baseline_from_arm.pyscript11.0 KB
tests/test_sweep_decode_capture.pyexecutablescript31.8 KB
tests/test_sweep_dp.pyexecutablescript45.4 KB
tests/test_sweep_fresh_cache.pyscript5.5 KB
tests/test_sweep_profiler_guard.pyscript22.6 KB
tests/test_sweep_real_prompts.pyscript20.1 KB
tests/test_sweep_refactor.pyexecutablescript37.6 KB
tests/test_sweep_round_slot.pyexecutablescript7.4 KB
tests/test_validation_report_phase_triage.pyexecutablescript10.0 KB
tests/test_validation_report_v2.pyexecutablescript6.2 KB
tests/test_verify_validation_gates.pyexecutablescript52.9 KB
tests/test_workload_matrix.pyexecutablescript19.3 KB

scripts/artifact_layout.json

View on GitHub
{
  "_comment": "Machine-readable AMMO V2 artifact-path manifest. Sole source of the PATH FACTS that references/artifact-layout.md states in prose and that both variants' layout warn hooks enforce. Consumers: .codex/hooks/ammo-artifact-layout-warn.sh and .codex/hooks/post_tool_use_guard.py both derive their allowlist from `slots[].path_template` + `placeholders`; neither restates a path. Judgment (disambiguation rules, prohibited-pattern rationale, resolution order) stays prose in artifact-layout.md. skills/ammo/tests/test_artifact_layout_manifest.py holds doc<->manifest parity.",
  "version": 1,
  "doc": "references/artifact-layout.md",
  "placeholders": {
    "{N}": "[0-9]+",
    "{D}": "[0-9]+",
    "{C}": "[0-9]+",
    "{BS}": "[0-9]+",
    "{op_id}": "[A-Za-z0-9_-]+",
    "{champion_id}": "[A-Za-z0-9_-]+",
    "{monitor_id}": "[A-Za-z0-9_-]+",
    "{stage}": "[A-Za-z0-9_*-]+",
    "{date}": "[0-9A-Za-z:_.-]+",
    "{slot}": "[A-Za-z0-9_-]+",
    "{timestamp}": "[0-9A-Za-z:_.-]+"
  },
  "_placeholders_comment": "Placeholder token -> regex fragment. A template segment holding no placeholder is matched literally. A trailing '/' makes the template a directory PREFIX (anything below it conforms); no trailing '/' makes it an exact file path. Consumers also accept every proper directory ancestor of a template, so a bare `rounds/3/sweeps` token is a directory reference, not drift.",
  "open_dirs": [
    "rounds/{N}/mining/",
    "rounds/{N}/audits/"
  ],
  "_open_dirs_comment": "Directories whose whole subtree conforms because the child filenames are agent-chosen, not a fixed set: the researcher adds mining evidence files, and the auditor writes per-cycle and per-track verdict filenames. The slots inside them stay enumerated below for their writer/gate/kind facts. Every other directory keeps its exact child set, so a track-root `validation_results_DRAFT.md` still warns.",
  "slots": {
    "state_json": {
      "path_template": "state.json",
      "writer": "orchestrator",
      "gate": null,
      "is_gate_slot": false,
      "kind": "authoritative",
      "notes": "Shape + enums owned by schemas/state.schema.json. Sub-agents never write it."
    },
    "state_lock": {
      "path_template": "state.json.lock",
      "writer": "ammo_state.py state_lock() + reconcile_track_state.py",
      "gate": null,
      "is_gate_slot": false,
      "kind": "diagnostic",
      "notes": "Empty flock file. Every writer of state.json serializes on it, so it sits next to state.json and is never deleted. Carries no campaign data."
    },
    "target_json": {
      "path_template": "target.json",
      "writer": "new_target.py + orchestrator (SHIP env promotion)",
      "gate": null,
      "is_gate_slot": false,
      "kind": "authoritative"
    },
    "report_md": {
      "path_template": "REPORT.md",
      "writer": "agents/ammo-report",
      "gate": null,
      "is_gate_slot": false,
      "kind": "authoritative",
      "notes": "Terminal deliverable, Stage 7b."
    },
    "report_assets": {
      "path_template": "report_assets/",
      "writer": "agents/ammo-report",
      "gate": null,
      "is_gate_slot": false,
      "kind": "derived",
      "notes": "5 required PNGs plus their gen_*.py generators."
    },
    "blockers": {
      "path_template": "blockers/{stage}_{date}.md",
      "writer": "orchestrator",
      "gate": null,
      "is_gate_slot": false,
      "kind": "authoritative",
      "notes": "Cross-round, NOT round-scoped."
    },
    "constraints": {
      "path_template": "rounds/{N}/constraints.md",
      "writer": "ammo-researcher",
      "gate": null,
      "is_gate_slot": false,
      "kind": "authoritative"
    },
    "validation_gate_report": {
      "path_template": "rounds/{N}/validation_gate_report.json",
      "writer": "verify_validation_gates.py --json-output",
      "gate": "cohort",
      "is_gate_slot": false,
      "kind": "authoritative",
      "notes": "One cohort-wide mechanical gate result per round."
    },
    "profiling_nsys": {
      "path_template": "rounds/{N}/profiling/nsys/",
      "writer": "run_vllm_bench_latency_sweep.py --nsys-profile",
      "gate": null,
      "is_gate_slot": false,
      "kind": "authoritative",
      "notes": "Attribution traces. Holds baseline_bs{BS}.nsys-rep plus opt/{op_id}/, integration/ and post_ship/ subdirs. Traces NEVER land inside a sweep output dir."
    },
    "profiling_ncu": {
      "path_template": "rounds/{N}/profiling/ncu/",
      "writer": "ammo-researcher",
      "gate": null,
      "is_gate_slot": false,
      "kind": "diagnostic",
      "notes": "Targeted hardware counters (sanity.csv + sanity_results.md). Its latency column is NOT a baseline."
    },
    "profiling_probe": {
      "path_template": "rounds/{N}/profiling/probe/",
      "writer": "legacy campaigns only",
      "gate": null,
      "is_gate_slot": false,
      "kind": "diagnostic",
      "notes": "Read-compatibility only. Not scaffolded, not a Stage 2 ranking input."
    },
    "profiling_torch": {
      "path_template": "rounds/{N}/profiling/torch_profile/",
      "writer": "legacy campaigns only",
      "gate": null,
      "is_gate_slot": false,
      "kind": "diagnostic",
      "notes": "Read-compatibility only."
    },
    "sweep_baseline": {
      "path_template": "rounds/{N}/sweeps/baseline/",
      "writer": "run_vllm_bench_latency_sweep.py --slot baseline",
      "gate": "stage_1",
      "is_gate_slot": false,
      "kind": "authoritative",
      "notes": "Clean Stage 1 E2E. Profiling flags forbidden. Holds json/golden_refs.json for Gate 5.1b."
    },
    "sweep_profiling": {
      "path_template": "rounds/{N}/sweeps/profiling/",
      "writer": "run_vllm_bench_latency_sweep.py --slot profiling",
      "gate": null,
      "is_gate_slot": false,
      "kind": "diagnostic",
      "notes": "Latency here is contaminated. Never a timing authority."
    },
    "sweep_profiling_prefill": {
      "path_template": "rounds/{N}/sweeps/profiling_prefill/",
      "writer": "run_vllm_bench_latency_sweep.py --slot profiling_prefill",
      "gate": null,
      "is_gate_slot": false,
      "kind": "diagnostic",
      "notes": "Contaminated latency AND shortened OSL, so any decode-shape claim needs a production-depth cross-check."
    },
    "sweep_opt": {
      "path_template": "rounds/{N}/sweeps/opt/{op_id}/",
      "writer": "run_vllm_bench_latency_sweep.py --slot opt/{op_id}",
      "gate": "5.3b",
      "is_gate_slot": true,
      "kind": "authoritative",
      "notes": "Clean per-track E2E timing authority. One per track per round; superseded runs archived."
    },
    "sweep_opt_correctness": {
      "path_template": "rounds/{N}/sweeps/opt_correctness/{op_id}/",
      "writer": "run_vllm_bench_latency_sweep.py --slot opt_correctness/{op_id}",
      "gate": "5.1b",
      "is_gate_slot": true,
      "kind": "authoritative",
      "notes": "Clean full-model correctness. Emits json/correctness_verdict.json."
    },
    "sweep_opt_profiling": {
      "path_template": "rounds/{N}/sweeps/opt_profiling/{op_id}/",
      "writer": "run_vllm_bench_latency_sweep.py --slot opt_profiling/{op_id}",
      "gate": "5.3a",
      "is_gate_slot": true,
      "kind": "authoritative",
      "notes": "Activation evidence only; latency is contaminated and never a timing authority."
    },
    "sweep_integration": {
      "path_template": "rounds/{N}/sweeps/integration/",
      "writer": "run_vllm_bench_latency_sweep.py --slot integration",
      "gate": "stage_6",
      "is_gate_slot": true,
      "kind": "authoritative",
      "notes": "Promoted baseline for round N+1 mining after a material SHIP, including when populated by the single-pass short-circuit."
    },
    "sweep_integration_profiling": {
      "path_template": "rounds/{N}/sweeps/integration_profiling/",
      "writer": "run_vllm_bench_latency_sweep.py --slot integration_profiling",
      "gate": "stage_6",
      "is_gate_slot": true,
      "kind": "authoritative",
      "notes": "Combined activation proof only; latency is contaminated."
    },
    "sweep_golden_capture": {
      "path_template": "rounds/{N}/sweeps/golden_capture/",
      "writer": "run_vllm_bench_latency_sweep.py --slot golden_capture",
      "gate": "post_ship",
      "is_gate_slot": false,
      "kind": "authoritative",
      "notes": "Emits json/golden_refs.json consumed by the next round's Gate 5.1b."
    },
    "sweep_post_ship_profiling": {
      "path_template": "rounds/{N}/sweeps/post_ship_profiling/",
      "writer": "run_vllm_bench_latency_sweep.py --slot post_ship_profiling",
      "gate": null,
      "is_gate_slot": false,
      "kind": "diagnostic",
      "notes": "Optional post-SHIP re-attribution capture; traces route to profiling/nsys/post_ship/. Contaminated latency, and T16 stays eliminated so it is never mandatory."
    },
    "mining_analysis": {
      "path_template": "rounds/{N}/mining/bottleneck_analysis.md",
      "writer": "ammo-researcher",
      "gate": "stage_2",
      "is_gate_slot": false,
      "kind": "authoritative",
      "notes": "Prose plus tables.md pasted verbatim. A hand-edited mined number is a BLOCKING audit finding."
    },
    "mining_config": {
      "path_template": "rounds/{N}/mining/mine_config.json",
      "writer": "ammo-researcher",
      "gate": null,
      "is_gate_slot": false,
      "kind": "authoritative",
      "notes": "Agent JUDGMENT only: segmentation policy, family labels + symbol patterns, trace-to-BS mapping, physical ceilings, prefill_active."
    },
    "mining_mined_json": {
      "path_template": "rounds/{N}/mining/mined.json",
      "writer": "mine_trace.py",
      "gate": "stage_2",
      "is_gate_slot": false,
      "kind": "authoritative",
      "notes": "Schema mine_trace/1. Sole writer of every mined number."
    },
    "mining_tables": {
      "path_template": "rounds/{N}/mining/tables.md",
      "writer": "mine_trace.py",
      "gate": null,
      "is_gate_slot": false,
      "kind": "derived",
      "notes": "The two mandated tables. Regenerated from mined.json; never hand-edited."
    },
    "debate_eligibility": {
      "path_template": "rounds/{N}/debate/eligibility_verdicts.md",
      "writer": "orchestrator (lead-owned)",
      "gate": null,
      "is_gate_slot": false,
      "kind": "authoritative",
      "notes": "CUT-4 four-gate verdict plus gated_revision. Write-once-then-append."
    },
    "debate_proposals": {
      "path_template": "rounds/{N}/debate/proposals/{champion_id}_proposal.md",
      "writer": "ammo-champion",
      "gate": null,
      "is_gate_slot": false,
      "kind": "authoritative",
      "notes": "Phase 0. In debate sub-round 1 the proposal IS the argument-of-record."
    },
    "debate_rounds": {
      "path_template": "rounds/{N}/debate/round_{D}/",
      "writer": "ammo-champion",
      "gate": null,
      "is_gate_slot": false,
      "kind": "authoritative",
      "notes": "D counts DEBATE sub-rounds, not campaign rounds. Holds {op_id}_{argument|critique_{target}|rebuttal}.md."
    },
    "debate_micro_experiments": {
      "path_template": "rounds/{N}/debate/micro_experiments/",
      "writer": "ammo-champion",
      "gate": null,
      "is_gate_slot": false,
      "kind": "diagnostic",
      "notes": "Feasibility scripts plus logs."
    },
    "debate_summary": {
      "path_template": "rounds/{N}/debate/summary.md",
      "writer": "orchestrator",
      "gate": null,
      "is_gate_slot": false,
      "kind": "derived",
      "notes": "Lead-authored human summary of the typed selected_candidates in state.json. Champions never write it; state.json wins any disagreement."
    },
    "track_validation_results": {
      "path_template": "rounds/{N}/tracks/{op_id}/validation_results.md",
      "writer": "generate_validation_report.py + agents/ammo-implementer",
      "gate": "T9",
      "is_gate_slot": false,
      "kind": "authoritative",
      "notes": "The ONLY authoritative file at the track root. check_projection_accuracy.py appends its section here."
    },
    "track_evidence": {
      "path_template": "rounds/{N}/tracks/{op_id}/evidence.json",
      "writer": "create_evidence_template.py + agents/ammo-implementer",
      "gate": null,
      "is_gate_slot": false,
      "kind": "authoritative",
      "notes": "Structured primary-evidence index; read by verify_validation_gates.py and reconcile_track_state.py."
    },
    "track_validation_summary": {
      "path_template": "rounds/{N}/tracks/{op_id}/validation_summary.json",
      "writer": "generate_validation_report.py",
      "gate": null,
      "is_gate_slot": false,
      "kind": "derived",
      "notes": "Machine summary of validation_results.md."
    },
    "track_diff_patch": {
      "path_template": "rounds/{N}/tracks/{op_id}/diff.patch",
      "writer": "agents/ammo-implementer",
      "gate": null,
      "is_gate_slot": false,
      "kind": "authoritative",
      "notes": "Refreshed from the merge base after each gate-relevant commit. Required for review, resume, integration and PR extraction."
    },
    "track_runtime_pkg_patch": {
      "path_template": "rounds/{N}/tracks/{op_id}/runtime_pkg.patch",
      "writer": "agents/ammo-implementer",
      "gate": null,
      "is_gate_slot": false,
      "kind": "authoritative",
      "notes": "Written ONLY when a materialized runtime was edited, derived from the roots worktree setup recorded."
    },
    "track_validator_tests": {
      "path_template": "rounds/{N}/tracks/{op_id}/validator_tests/",
      "writer": "agents/ammo-implementer",
      "gate": "5.1a+5.2",
      "is_gate_slot": false,
      "kind": "authoritative",
      "notes": "test_correctness.py, bench_gate_5_2.py, gate_5_1a_results.json, gate_5_2_results.json."
    },
    "track_monitor_audits": {
      "path_template": "rounds/{N}/tracks/{op_id}/monitor_audits/",
      "writer": "ammo-transcript-monitor",
      "gate": null,
      "is_gate_slot": false,
      "kind": "diagnostic",
      "notes": "Impl-stage only. Monitors write here, never to the campaign root. Debate champions have no monitors, so there is no debate/monitor_audits/."
    },
    "track_scratch": {
      "path_template": "rounds/{N}/tracks/{op_id}/_scratch/",
      "writer": "agents/ammo-implementer",
      "gate": null,
      "is_gate_slot": false,
      "kind": "diagnostic",
      "notes": "All drafts, debug scripts and iteration logs. Drafts look like finals, so nothing here is ever cited as a verdict."
    },
    "audit_stage_1": {
      "path_template": "rounds/{N}/audits/stage_1.md",
      "writer": "ammo-auditor",
      "gate": "T_AUDIT_S1",
      "is_gate_slot": false,
      "kind": "authoritative",
      "notes": "Verdict for T_AUDIT_S1. The stage transition reads the matching audit.stage_1.passed_at stamp."
    },
    "audit_stage_2": {
      "path_template": "rounds/{N}/audits/stage_2.md",
      "writer": "ammo-auditor",
      "gate": "T_AUDIT_S2",
      "is_gate_slot": false,
      "kind": "authoritative",
      "notes": "Verdict for T_AUDIT_S2. The stage transition reads the matching audit.stage_2.passed_at stamp."
    },
    "audit_stage_45": {
      "path_template": "rounds/{N}/audits/stage_45.md",
      "writer": "ammo-auditor",
      "gate": "T_AUDIT_S45",
      "is_gate_slot": false,
      "kind": "authoritative",
      "notes": "Verdict for T_AUDIT_S45. The stage transition reads the matching audit.stage_45.passed_at stamp."
    },
    "audit_stage_67": {
      "path_template": "rounds/{N}/audits/stage_67.md",
      "writer": "ammo-auditor",
      "gate": "T_AUDIT_S67",
      "is_gate_slot": false,
      "kind": "authoritative",
      "notes": "Verdict for T_AUDIT_S67. The stage transition reads the matching audit.stage_67.passed_at stamp."
    },
    "audit_stage45_partial": {
      "path_template": "rounds/{N}/audits/stage_45_partial_{op_id}.md",
      "writer": "ammo-delegate dispatched by the orchestrator",
      "gate": null,
      "is_gate_slot": false,
      "kind": "diagnostic",
      "notes": "Optional early per-track evidence, NOT a verdict. Consumed by the real auditor at T_AUDIT_S45."
    },
    "audit_cycle": {
      "path_template": "rounds/{N}/audits/stage_{stage}_cycle_{C}.md",
      "writer": "ammo-auditor",
      "gate": "T_AUDIT",
      "is_gate_slot": false,
      "kind": "authoritative",
      "notes": "Re-audit after BLOCKED (C=2,3)."
    },
    "archive": {
      "path_template": "rounds/{N}/_archive/{slot}_{timestamp}/",
      "writer": "run_vllm_bench_latency_sweep.py",
      "gate": null,
      "is_gate_slot": false,
      "kind": "diagnostic",
      "notes": "Quarantined superseded runs. The ONLY place a timestamp may appear in a directory name."
    }
  }
}
slack-gif-creatoranthropics/skillsKnowledge and utilities for creating animated GIFs optimized for Slack. Provides constraints, validation tools, and animation concepts. Use when users request animated GIFs for Slack like "make me a GIF of X doing Y for Slack."Stars: 198kUpdated Claude CodeCodexVerifiedInstall
ponytailDietrichGebert/ponytailLazy senior dev mode for any coding task: the smallest change that fully solves it, and a reply a busy human understands in one read.Stars: 194kUpdated Claude CodeCodexVerifiedInstall
ponytail-debtDietrichGebert/ponytailHarvest every shortcut: comment into one debt ledger, so deferrals get tracked instead of forgotten. One-shot report.Stars: 194kUpdated Claude CodeCodexVerifiedInstall
ponytail-debtDietrichGebert/ponytailList every `shortcut:` comment (and older `ponytail:` ones) as a debt ledger. One-shot report, changes nothing. Use for "ponytail debt", "what did ponytail defer", "list the shortcuts", /ponytail-debt.Stars: 194kUpdated Claude CodeCodexVerifiedInstall