Skip to content

hashgraph-online/awesome-codex-plugins / plugins/beastlabai/multi-llm-plugin/skills/multi-llm

multi-llm

Flagged

Orchestrate plan reviews, task generation, implementation and code reviews across several LLM CLIs in parallel, or ask every model one free-text question about a plan. Use when the user says: review my plan, multi-llm, run a code review, generate tasks from a plan, implement this plan, ask the models about my plan.

FileKindSize
SKILL.mdskill md4.0 KB
apply_code_fixes_orchestrator.pyscript56.6 KB
apply_commit_fixes_orchestrator.pyscript2.3 KB
apply_suggestions_orchestrator.pyscript4.5 KB
apply_task_suggestions_orchestrator.pyscript19.1 KB
ask_orchestrator.pyscript162.5 KB
assets/icons/Anthropic.svgother584 B
assets/icons/Codex.svgother1.3 KB
assets/icons/Copilot.svgother1.5 KB
assets/icons/Cursor.svgother793 B
assets/icons/GoogleGemini.svgother623 B
assets/icons/Goose.svgother3.2 KB
assets/icons/Kilocode.svgother531 B
assets/icons/MiniMax.pngbinary1.5 KB
assets/icons/Mistral.svgother717 B
assets/icons/MoonshotAI.svgother957 B
assets/icons/Nvidia.svgother2.4 KB
assets/icons/OhMyPi.svgother434 B
assets/icons/OpenAI.svgother1.7 KB
assets/icons/Opencode.svgother325 B
assets/icons/Pi.svgother440 B
assets/icons/Qwen.svgother1.7 KB
assets/icons/Zhipu.pngbinary442 B
assets/icons/xAI.svgother343 B
assets/omp/overlay.ymlconfig2.4 KB
check_workflow_prerequisites.pyscript74.5 KB
code_review_orchestrator.pyscript324.0 KB
display_decisions.pyscript9.8 KB
finalize_tracking.pyscript1.3 KB
full_workflow_orchestrator.pyscript58.8 KB
generate_tasks_orchestrator.pyscript44.2 KB
hosts/capabilities.yamlconfig49.9 KB
hosts/codex/plugin-skill.md.tmplother2.8 KB
hosts/codex/shim.md.tmplother7.3 KB
hosts/opencode/plugin-command.md.tmplother2.2 KB
hosts/opencode/shim.md.tmplother6.8 KB
hosts/render_capabilities.pyscript48.4 KB
hosts/snippets/poll-loop.mdmarkdown4.4 KB
hosts/snippets/uv-preflight.mdmarkdown2.2 KB
hosts/uv-missing.txtmarkdown1.5 KB
implement_orchestrator.pyscript156.0 KB
improve_code_orchestrator.pyscript1.6 KB
improve_commits_orchestrator.pyscript2.5 KB
improve_plan_orchestrator.pyscript1.6 KB
init_config.pyscript57.6 KB
instructions/apply-code-fixes.mdmarkdown52.0 KB
instructions/apply-commit-fixes.mdmarkdown10.3 KB
instructions/apply-suggestions.mdmarkdown48.2 KB
instructions/apply-task-suggestions.mdmarkdown53.9 KB
instructions/ask.mdmarkdown27.6 KB
instructions/full-workflow.mdmarkdown24.6 KB
instructions/generate-tasks.mdmarkdown18.1 KB
instructions/implement.mdmarkdown54.0 KB
instructions/improve-code.mdmarkdown25.0 KB
instructions/improve-commits.mdmarkdown24.8 KB
instructions/improve-plan.mdmarkdown26.2 KB
instructions/init-config.mdmarkdown6.6 KB
instructions/review-code.mdmarkdown43.5 KB
instructions/review-commits.mdmarkdown11.4 KB
instructions/review-plan.mdmarkdown38.4 KB
instructions/review-tasks.mdmarkdown37.8 KB
modes.pyscript58.8 KB
multi_llm.pyscript101.1 KB
prompts/ask.txtmarkdown948 B
prompts/code_review.txtmarkdown4.5 KB
prompts/commit_review.txtmarkdown6.2 KB
prompts/consolidate_suggestions.txtmarkdown3.1 KB
prompts/implementation_task.txtmarkdown1.4 KB
prompts/plan_review.txtmarkdown8.0 KB
prompts/task_review.txtmarkdown8.3 KB
providers.yamlconfig19.6 KB
pyproject.tomlmanifest6.8 KB
references/apply-write-executor.mdmarkdown3.0 KB
references/consolidation.mdmarkdown5.6 KB
references/host-capabilities.mdmarkdown13.5 KB
references/human-decision-batch.mdmarkdown17.2 KB
references/implement-prerequisites.mdmarkdown6.0 KB
references/implement-write-executor.mdmarkdown4.9 KB
references/interrupted-run-recovery.mdmarkdown3.1 KB
references/own-check.mdmarkdown4.3 KB
references/resume-detection.mdmarkdown8.5 KB
references/revalidation.mdmarkdown6.9 KB
references/review-code-apply-fixes.mdmarkdown4.3 KB
references/salvage-handling.mdmarkdown12.5 KB
references/summary-format.mdmarkdown10.0 KB
references/uv-check.mdmarkdown5.7 KB
references/wave-batching.mdmarkdown5.2 KB
review_commits_orchestrator.pyscript2.3 KB
review_plan_orchestrator.pyexecutablescript22.3 KB
review_tasks_orchestrator.pyscript10.7 KB
schemas/cascade_record.schema.jsonconfig6.6 KB
schemas/code_review_issues.schema.jsonconfig2.0 KB
schemas/consolidated.schema.jsonconfig6.8 KB
schemas/consolidation_batch.schema.jsonconfig3.7 KB
schemas/generated_tasks.schema.jsonconfig3.1 KB
schemas/host_capabilities.schema.jsonconfig38.3 KB
schemas/human_decision_verdicts.schema.jsonconfig2.6 KB
schemas/result_line.schema.jsonconfig10.1 KB
schemas/review_suggestions.schema.jsonconfig2.4 KB
schemas/state_file.schema.jsonconfig21.5 KB
schemas/status_file.schema.jsonconfig24.9 KB
schemas/status_json.schema.jsonconfig10.8 KB
schemas/subagent_tasks.schema.jsonconfig20.5 KB
schemas/task_decomposition.schema.jsonconfig1.2 KB
schemas/write_batch_result.schema.jsonconfig8.8 KB
templates/config/providers.override.yamlconfig17.2 KB
templates/consolidated_report_template.htmlother44.1 KB
templates/pr_report_template.htmlother181.2 KB
templates/report_template.htmlother67.5 KB
templates/shared/report_core.cssother4.1 KB
templates/shared/report_core.jsscript25.5 KB
update_plan_tasks.pyscript20.9 KB
utils/__init__.pyscript284 B
utils/advisory.pyscript4.8 KB
utils/apply_orchestrator_base.pyscript220.6 KB
utils/apply_output_helpers.pyscript10.9 KB
utils/apply_path_helpers.pyscript683 B
utils/apply_report.pyscript51.2 KB
utils/apply_selection_helpers.pyscript6.8 KB
utils/apply_suggestion_phase.pyscript11.7 KB
utils/atomic_io.pyscript12.9 KB
utils/attempt_classifier.pyscript31.4 KB
utils/backup.pyscript2.0 KB
utils/batch_done.pyscript79.2 KB
utils/batch_ledger.pyscript27.4 KB
utils/batch_prompt.pyscript29.3 KB
utils/batch_prompt_builder.pyscript12.2 KB
utils/batch_reply.pyscript32.9 KB
utils/cascade_archive.pyscript50.6 KB
utils/cascade_ledger.pyscript39.3 KB
utils/cascade_provenance.pyscript22.0 KB
utils/check_records.pyscript59.3 KB
utils/check_reply.pyscript9.5 KB
utils/check_tree.pyscript1.2 KB
utils/cli_common.pyscript91.2 KB
utils/cli_version_cache.pyscript14.0 KB
utils/clock.pyscript3.5 KB
utils/code_fix_batcher.pyscript30.6 KB
utils/codex_trust.pyscript12.9 KB
utils/commit_apply.pyscript73.3 KB
utils/commit_attribution.pyscript21.2 KB
utils/commit_brief.pyscript32.2 KB
utils/commit_front_door.pyscript81.6 KB
utils/commit_paths.pyscript18.0 KB
utils/commit_range.pyscript61.2 KB
utils/commit_report.pyscript19.9 KB
utils/commit_review.pyscript413 B
utils/commit_spelling.pyscript7.9 KB
utils/commit_write_guard.pyscript9.9 KB
utils/concurrency.pyscript11.6 KB
utils/consolidation.pyscript87.3 KB
utils/detach.pyscript186.8 KB
utils/dispatch_spawn.pyscript56.3 KB
utils/doctor.pyscript76.7 KB
utils/env_vars.pyscript20.5 KB
utils/failure_capture.pyscript10.1 KB
utils/file_discovery.pyscript11.0 KB
utils/filtering.pyscript19.9 KB
utils/git_diff.pyscript23.6 KB
utils/git_utils.pyscript86.6 KB
utils/held_folder.pyscript21.3 KB
utils/host_capabilities.pyscript48.0 KB
utils/host_install.pyscript77.6 KB
utils/host_paths.pyscript12.0 KB
utils/host_routes/__init__.pyscript5.7 KB
utils/host_routes/codex.pyscript18.3 KB
utils/host_routes/npm.pyscript12.2 KB
utils/host_shim.pyscript14.7 KB
utils/html_report_generator.pyscript118.9 KB
utils/human_decisions.pyscript38.3 KB
utils/human_task_plan.pyscript30.2 KB
utils/importance.pyscript6.2 KB
utils/improve_cascade.pyscript63.7 KB
utils/improve_cascade_next.pyscript22.2 KB
utils/improve_steps.pyscript32.4 KB
utils/improve_workflow.pyscript114.4 KB
utils/interactive.pyscript31.6 KB
utils/json_extractor.pyscript61.5 KB
utils/judge_vocab.pyscript5.3 KB
utils/known_checks.pyscript37.5 KB
utils/landlock_dispatch.pyscript103.8 KB
utils/landlock_exec.pyscript21.4 KB
utils/launch_limiter.pyscript13.5 KB
utils/llm_client.pyscript126.9 KB
utils/manual_tasks.pyscript14.9 KB
utils/metrics.pyscript68.1 KB
utils/next_phase.pyscript96.1 KB
utils/next_phase_words.pyscript6.4 KB
utils/output_handler.pyscript23.2 KB
utils/phase_walk.pyscript6.4 KB
utils/plan_updater.pyscript17.9 KB
utils/positionals.pyscript5.8 KB
utils/prior_changes.pyscript28.7 KB
utils/probe_state.pyscript17.9 KB
utils/process_capture.pyscript37.6 KB
utils/process_control.pyscript228.7 KB
utils/process_state.pyscript6.6 KB
utils/progress_tracker.pyscript92.2 KB
utils/prompt_channels.pyscript34.1 KB
utils/prompt_files.pyscript24.5 KB

schemas/subagent_tasks.schema.json

View on GitHub
{
  "$schema": "http://json-schema.org/draft-07/schema#",
  "$id": "subagent_tasks.schema.json",
  "title": "Materialized subagent task file (tasks_schema_version 2)",
  "description": "Validates every materialized task file the orchestrators emit — validation_tasks.json, revalidation_tasks.json, consolidation_tasks.json, salvage_task_<model>.json, human_decision_tasks.json and task_generation_tasks.json. Version 1 is the historical reference-based shape that carried no prompt; a reader that finds tasks_schema_version missing or != 2 refuses the file with exit 3 and re-materializes the phase (compatibility policy is regenerate, never upgrade-in-place). This key is deliberately distinct from metadata.schema_version, which versions validation *results*.",
  "type": "object",
  "required": ["tasks_schema_version", "kind", "inputs"],
  "anyOf": [
    {"required": ["tasks"]},
    {"required": ["batches"]}
  ],
  "properties": {
    "tasks_schema_version": {
      "const": 2,
      "description": "Integer task-file format version. Missing means 1."
    },
    "context_contract": {
      "type": "integer",
      "minimum": 1,
      "description": "The validation context contract the file was materialized under (validation/revalidation files). 2 = over-limit documents named by hash-bound snapshot path; missing means 1, whose over-limit prompts were cut at the inline limit and are retired on resume/reaggregate."
    },
    "kind": {
      "$ref": "#/$defs/kind",
      "description": "The kind every task in this file has."
    },
    "inputs": {
      "$ref": "#/$defs/inputs",
      "description": "File-level copy of the inputs block; identical to every task's own inputs."
    },
    "round_id": {
      "type": "string",
      "pattern": "^[0-9a-f]{16,64}$",
      "description": "Validation task files of the review phases: the review round `.status.json` recorded when the file was materialized (utils/review_round.py), so a resume and review-code's partial-completion guard can tell this round's file from an earlier round's. Absent on a file written before round ids, or with no round on record."
    },
    "context_plan_file": {
      "type": "string",
      "description": "Frozen context-plan snapshot, relative to the phase directory (absolute across drives). Its raw bytes must match inputs.context_plan_sha256 before dispatch."
    },
    "tasks": {
      "type": "array",
      "items": {"$ref": "#/$defs/task"},
      "description": "The task list. Batched kinds carry it under the historical name `batches` instead, so the list is never duplicated on disk; use one accessor (utils.subagent_tasks.tasks_of) rather than guessing."
    },
    "batches": {
      "type": "array",
      "items": {"$ref": "#/$defs/task"},
      "description": "The task list under its historical name, for the batched kinds whose merge and reaggregation paths already read it."
    }
  },
  "$defs": {
    "kind": {
      "enum": [
        "validation",
        "revalidation",
        "consolidation",
        "salvage",
        "human_decision_judge",
        "task_generation",
        "apply_suggestions",
        "apply_task_suggestions",
        "apply_code_fixes",
        "implement"
      ]
    },
    "inputs": {
      "type": "object",
      "additionalProperties": false,
      "required": ["plan_sha256"],
      "properties": {
        "plan_sha256": {
          "type": "string",
          "description": "Content hash of the plan file, computed the same way StateManager._compute_plan_hash does (sha256 hexdigest truncated to 16 chars). Empty string when the phase has no plan file."
        },
        "grouped_sha256": {
          "type": "string",
          "description": "Content hash of grouped.json."
        },
        "tasks_sha256": {
          "type": "string",
          "description": "Content hash of the task list under review \u2014 review-tasks only, whose validation prompts embed the tasks text beside the plan."
        },
        "diff_base": {
          "type": "string",
          "description": "Git ref the diff was taken against (state.head_at_start) — code-review and apply phases only."
        },
        "diff_sha256": {
          "type": "string",
          "description": "Content hash of the diff text — code-review and apply phases only."
        },
        "reviewed_code_sha256": {
          "type": "string",
          "description": "Content hash of the code excerpt the prompts embed — code-review only. Moves when a reviewed file is edited between fan-out and validation, which diff_base alone cannot see."
        },
        "context_plan_sha256": {
          "type": "string",
          "description": "SHA-256 of a commit review's frozen context-plan snapshot, as commit_range.plan_sha256 recorded it — code-review on a commit brief with --context-plan only. A resume or reaggregate whose snapshot no longer hashes to it is refused."
        },
        "salvage_request_sha256": {
          "type": "string",
          "description": "Content hash of the salvage request record — salvage only."
        },
        "items_sha256": {
          "type": "string",
          "description": "Content hash of the judged item set — human_decision_judge only."
        },
        "prior_rounds_sha256": {
          "type": "string",
          "description": "SHA-256 of the cascade stage's frozen validator ledger (ledger-validator.md) that every batch prompt embeds inline — validation/revalidation of a cascade stage whose ledger is not empty only; equal to that file's sha256 in the stage's manifest. Absent everywhere else."
        },
        "cascade_stage": {
          "type": "string",
          "description": "<cascade_id>:<stage> of the cascade stage whose items these judge tasks decide — human_decision_judge of a cascade stage only. A verdict file beside a task file that does not name the current stage (or its judge_ledger_sha256) is moved aside before any verdict is read. Absent everywhere else."
        },
        "judge_ledger_sha256": {
          "type": "string",
          "description": "SHA-256 of the cascade stage's frozen judge ledger (ledger-judge.md) the prompts embed — human_decision_judge of a cascade stage only, beside cascade_stage. Absent everywhere else."
        }
      }
    },
    "task": {
      "type": "object",
      "additionalProperties": false,
      "required": [
        "tasks_schema_version",
        "task_id",
        "kind",
        "prompt",
        "output_name",
        "result_channel",
        "context_mode",
        "capability_tier",
        "executor_role",
        "inputs"
      ],
      "properties": {
        "tasks_schema_version": {"const": 2},
        "task_id": {
          "type": "string",
          "minLength": 1,
          "description": "Stable, path-free identifier for this task. Named in every refusal message."
        },
        "kind": {"$ref": "#/$defs/kind"},
        "prompt": {
          "type": "string",
          "minLength": 1,
          "description": "The complete prompt. Under context_mode inline it embeds every input the subagent needs; it never names plan_file, grouped_file or output_path."
        },
        "output_name": {
          "type": "string",
          "minLength": 1,
          "pattern": "^[A-Za-z0-9._-]+$",
          "description": "Bare basename resolved against the phase dir the task file was loaded from."
        },
        "output_path": {
          "type": "string",
          "description": "Derived absolute convenience path for native invokers. Invoker plumbing, never a subagent instruction."
        },
        "prompt_path": {
          "type": "string",
          "description": "Optional absolute path to a sidecar holding exactly the bytes of `prompt` (validation, revalidation, consolidation and task generation only). Invoker plumbing, never a subagent instruction: a host whose subagent tool takes a prompt string may hand the subagent this file instead of re-emitting its prompt (~65 KB for a validation batch, ~21 KB for task generation) through its own context. `prompt` stays the source of truth and every other reader ignores this key."
        },
        "prompt_lines": {
          "type": "integer",
          "minimum": 1,
          "description": "Line count of `prompt`, and therefore of the sidecar, so a host can tell a subagent to read through the last line rather than trust a default page size."
        },
        "result_channel": {
          "enum": ["file", "inline-json"],
          "description": "Producer intent. Every materialized task is `file`. `inline-json` names only the legacy Claude-Code-interactive shape and is refused before dispatch with exit 3."
        },
        "context_mode": {
          "enum": ["inline", "workspace-read"],
          "description": "How the task's inputs travel. `inline` embeds content in prompt and runs against any runner. `workspace-read` references absolute paths and requires a runner declaring supports_workspace_reads."
        },
        "capability_tier": {
          "enum": ["full", "cheap"],
          "description": "Role annotation, never a model. Resolution chain: model_hint (when set and not `auto`) -> capability_tier -> driver.subagent_runner_cheap / driver.subagent_runner -> host default. `cheap` is a permission to downgrade, and the per-kind rules below forbid it wherever AGENTS.md forbids downgrading. No currently materialized kind permits it: `cheap` is reserved for the human-decision fan-out coordinator, a dispatch-only role that writes no verdict and therefore has no task file."
        },
        "executor_role": {
          "enum": ["agent", "human"],
          "description": "`agent` means any invoker may dispatch this. `human` means no invoker ever dispatches it; it routes to the human-task pause and the executor refuses it before dispatch (exit 3)."
        },
        "inputs": {"$ref": "#/$defs/inputs"},
        "model_hint": {
          "type": "string",
          "description": "Free-form model string, or `auto`. Keeps its historical meaning; `full`/`cheap` belong to capability_tier alone and are a schema error here.",
          "not": {"enum": ["full", "cheap"]}
        },
        "subagent_type": {
          "enum": ["general-purpose", "human"],
          "description": "Optional Claude-Code-only metadata, demoted for one release. No reader branches on it. Must agree with executor_role."
        },
        "context_sources": {
          "type": "array",
          "description": "Validation/revalidation only, and only when some context block is not inline (plans/validation-context-by-path-and-codex-0156-fallout.md \u00a72.3 step 7). One record per context document. A `path` record is the hash-bound snapshot the prompt names by `abs_path`; `path` is relative to the phase dir when the snapshot lives there, so a moved phase dir can rebind it after re-verifying `sha256` (raw bytes). A `truncated` record says why no file is named.",
          "items": {
            "type": "object",
            "required": ["label", "mode"],
            "properties": {
              "label": {"enum": ["plan", "tasks", "reviewed-code", "diff"]},
              "kind": {"enum": ["plan", "tasks", "reviewed-code", "diff"]},
              "mode": {"enum": ["inline", "path", "truncated"]},
              "path": {"type": ["string", "null"]},
              "abs_path": {"type": ["string", "null"]},
              "in_phase_dir": {"type": "boolean"},
              "sha256": {"type": ["string", "null"]},
              "bytes": {"type": "integer", "minimum": 0},
              "chars_shown": {"type": "integer", "minimum": 0},
              "truncated": {"type": "boolean"},
              "sections_total": {"type": "integer", "minimum": 0},
              "sections_shown": {"type": "integer", "minimum": 0},
              "dropped_sections": {"type": "array", "items": {"type": "string"}},
              "dropped_files": {"type": "array", "items": {"type": "string"}},
              "reason": {"type": ["string", "null"]}
            }
          }
        },
        "plan_file": {"type": "string", "description": "Provenance only."},
        "grouped_file": {"type": "string", "description": "Provenance only."},
        "batch_index": {"type": "integer", "minimum": 0},
        "group_ids": {"type": "array", "items": {"type": "string"}},
        "group_indices": {"type": "array", "items": {"type": "integer"}},
        "groups_count": {"type": "integer", "minimum": 0},
        "is_high_priority": {"type": "boolean"},
        "section_key": {"type": "string"},
        "group_id": {"type": "string", "minLength": 1},
        "items_count": {"type": "integer", "minimum": 0},
        "suggestion_id": {"type": "string", "minLength": 1},
        "model": {"type": "string"},
        "phase": {"type": "string"},
        "repo_root": {"type": "string"},
        "declared_files": {
          "type": "array",
          "items": {"type": "string"},
          "description": "WRITE kinds only. The file set this batch declares it will touch, repo-root-relative. Drives D3h's pairwise-disjoint co-scheduling admission and D3l's pre-dispatch write-policy check; an EMPTY list means 'undeclared', which runs the task alone rather than being treated as 'touches nothing'."
        },
        "snapshot_files": {
          "type": "array",
          "items": {"type": "string"},
          "description": "WRITE kinds only. Files the executor snapshots under {phase}/pre_batch/{task_id}/ before dispatch, so the post-run byte comparison (D3g) and `util restore-batch` (D3l) have a rollback point. Defaults to declared_files when absent."
        },
        "expected_ids": {
          "type": "array",
          "items": {"type": "string"},
          "description": "WRITE kinds only. The suggestion/group/task ids this batch is supposed to apply. D3g spot-checks that THESE landed rather than merely that bytes moved: a result reporting only some of them with a non-empty diff is a partial application, not a success."
        },
        "prior_changes_id": {
          "type": "string",
          "minLength": 1,
          "description": "WRITE kinds only. The stable `--id` for `prior_changes.py append` (suggestion/group/section/task key — never the loop index), so a re-dispatched batch appends once."
        },
        "can_parallelize": {
          "type": "boolean",
          "description": "WRITE kinds only. The batcher's opinion that this task may run beside its batch siblings. Necessary but not sufficient: driver.max_concurrent_write_subagents > 1 and the disjointness check must also agree (D3h)."
        },
        "depends_on": {
          "type": "array",
          "items": {"type": "string"},
          "description": "WRITE kinds only (implement). Task ids this task depends on; the human-task strategies walk these edges (D3n)."
        },
        "is_human": {
          "type": "boolean",
          "description": "WRITE kinds only. Mirror of `executor_role == \"human\"` for readers that predate D2h. The schema requires the two to agree."
        },
        "file_key": {
          "type": "string",
          "description": "WRITE kinds only (apply_code_fixes). The file a code-fix batch is partitioned by (utils/code_fix_batcher.py). Provenance for the declared file set, never a second source of truth."
        },
        "task_title": {
          "type": "string",
          "description": "WRITE kinds only. Human-readable label for logs, metrics and the implementation summary."
        },
        "scope": {
          "type": "object",
          "description": "Apply WRITE kinds under --scope-policy preserve only (utils/scope_apply.task_block). Absent means the task was materialized under expand (legacy tasks too). The executor's dispatch and native util write-prompt refuse a task whose policy, baseline or binding does not match the round's effective policy, and render the preserve writer guidance from it.",
          "required": ["policy", "baseline_id", "binding", "items"],
          "properties": {
            "policy": {"enum": ["expand", "preserve"]},
            "policy_source": {"type": "string"},
            "baseline_id": {"type": ["string", "null"], "description": "The review round's host-computed scope baseline id."},
            "binding": {"type": "string", "description": "sha256 over policy, baseline_id and items."},
            "items": {
              "type": "array",
              "items": {
                "type": "object",
                "required": ["id", "relation"],
                "properties": {
                  "id": {"type": "string", "description": "The member id, as in expected_ids."},
                  "group_id": {"type": "string"},
                  "relation": {"enum": ["in_scope", "scope_extension", "unclear"]},
                  "reason": {"type": "string"},
                  "proposal_digest": {"type": ["string", "null"], "description": "The validated proposal digest the verdict was bound to."},
                  "selected_remedy": {"type": "string", "description": "The remedy validation selected; the item's instruction."}
                }
              }
            }
          }
        }
      },
      "allOf": [
        {
          "if": {"required": ["kind"], "properties": {"kind": {"enum": ["validation", "revalidation"]}}},
          "then": {
            "required": ["batch_index", "group_ids", "capability_tier"],
            "properties": {
              "capability_tier": {"const": "full"},
              "context_mode": {"enum": ["inline", "workspace-read"]},
              "inputs": {"required": ["plan_sha256", "grouped_sha256"]}
            }
          }
        },
        {
          "if": {"required": ["kind"], "properties": {"kind": {"not": {"enum": ["validation", "revalidation"]}}}},
          "then": {"not": {"required": ["context_sources"]}}
        },
        {
          "if": {"required": ["kind"], "properties": {"kind": {"const": "consolidation"}}},
          "then": {
            "required": ["batch_index", "group_ids"],
            "properties": {
              "capability_tier": {"const": "full"},
              "context_mode": {"const": "inline"},
              "inputs": {"required": ["plan_sha256", "grouped_sha256"]}
            }
          }
        },
        {
          "if": {"required": ["kind"], "properties": {"kind": {"const": "salvage"}}},
          "then": {
            "required": ["suggestion_id"],
            "properties": {
              "capability_tier": {"const": "full"},
              "inputs": {"required": ["salvage_request_sha256"]}
            }
          }
        },
        {
          "if": {"required": ["kind"], "properties": {"kind": {"const": "human_decision_judge"}}},
          "then": {
            "required": ["group_id"],
            "properties": {
              "capability_tier": {"const": "full"},
              "context_mode": {"const": "inline"}
            }
          }
        },
        {
          "if": {"required": ["kind"], "properties": {"kind": {"const": "task_generation"}}},
          "then": {
            "required": ["repo_root"],
            "properties": {
              "capability_tier": {"const": "full"},
              "context_mode": {"const": "workspace-read"}
            }
          }
        },
        {
          "if": {
            "required": ["kind"],
            "properties": {
              "kind": {
                "enum": [
                  "apply_suggestions",
                  "apply_task_suggestions",
                  "apply_code_fixes",
                  "implement"
                ]
              }
            }
          },
          "then": {
            "required": ["batch_index", "declared_files", "prior_changes_id"],
            "properties": {
              "capability_tier": {"const": "full"},
              "context_mode": {"const": "workspace-read"}
            }
          }
        },
        {
          "if": {"required": ["kind"], "properties": {"kind": {"not": {"enum": ["validation", "revalidation", "consolidation", "task_generation"]}}}},
          "then": {"not": {"anyOf": [{"required": ["prompt_path"]}, {"required": ["prompt_lines"]}]}}
        },
        {
          "if": {"required": ["is_human"], "properties": {"is_human": {"const": true}}},
          "then": {"properties": {"executor_role": {"const": "human"}}}
        },
        {
          "if": {"required": ["is_human"], "properties": {"is_human": {"const": false}}},
          "then": {"properties": {"executor_role": {"const": "agent"}}}
        },
        {
          "if": {"required": ["executor_role"], "properties": {"executor_role": {"const": "human"}}},
          "then": {"properties": {"subagent_type": {"const": "human"}}}
        },
        {
          "if": {"required": ["executor_role"], "properties": {"executor_role": {"const": "agent"}}},
          "then": {"properties": {"subagent_type": {"const": "general-purpose"}}}
        }
      ]
    }
  }
}
receiving-code-reviewobra/superpowersUse when receiving code review feedback, before implementing suggestions, especially if feedback seems unclear or technically questionable - requires technical rigor and verification, not performative agreement or blind implementationStars: 322kUpdated Claude CodeCodexVerifiedInstall
requesting-code-reviewobra/superpowersUse when completing tasks, implementing major features, or before merging to verify work meets requirementsStars: 322kUpdated Claude CodeCodexVerifiedInstall
using-git-worktreesobra/superpowersUse when starting feature work that needs isolation from current workspace or before executing implementation plans - ensures an isolated workspace exists via native tools or git worktree fallbackStars: 322kUpdated Claude CodeCodexVerifiedInstall
verification-before-completionobra/superpowersUse when about to claim work is complete, fixed, or passing, before committing or creating PRs - requires running verification commands and confirming output before making any success claims; evidence before assertions alwaysStars: 322kUpdated Claude CodeCodexVerifiedInstall