j-mauro-r/SAT-Dengue-V2 / .agents/skills/deepworkplan
deepworkplan
VerifiedDeepWorkPlan — turn any repo AI-first and run Deep Work Plans. Routes to create, execute, refine, resume, status, verify, upgrade, and repo-onboarding sub-skills based on intent. Use when the developer wants to plan, execute, manage, or verify structured multi-task work, or make a repository AI-agent-ready.
spec/schema/plan-contract-v7.schema.json
View on GitHub{
"$schema": "https://json-schema.org/draft/2020-12/schema",
"$id": "https://deepworkplan.com/schema/plan-contract/v7.json",
"description": "Deep Work Plan outcome and authority contract, v7 generation (spec/V7_CONTRACT.md). A strict superset of plan-contract/v6: every v6 rule, identity and revision semantic applies unchanged (spec/V6_CONTRACT.md); the only addition is the optional boolean tasks[].parallel_safe, the create-time marker that makes a task eligible for delegation when the contract also grants agent_delegation. Generation is detected by this schema URL; a v6 plan never becomes v7 and the v6 schema bytes are unchanged.",
"type": "object",
"additionalProperties": false,
"required": [
"schema",
"spec_version",
"plan",
"revision",
"created_at",
"outcome",
"acceptance",
"invariants",
"scope",
"authorization",
"permissions",
"dependencies",
"resource_envelope",
"tasks"
],
"properties": {
"schema": {
"const": "https://deepworkplan.com/schema/plan-contract/v7.json"
},
"spec_version": {
"type": "string",
"pattern": "^\\d+\\.\\d+\\.\\d+$"
},
"plan": {
"type": "string",
"description": "Plan folder name: optional monotonic numeric ID of at least three digits, followed by a 2–5-word lowercase snake_case slug; legacy unnumbered names remain valid.",
"pattern": "^PLAN_([0-9]{3,}_)?[a-z0-9]+(_[a-z0-9]+){1,4}$"
},
"revision": {
"type": "integer",
"minimum": 1
},
"contract_id": {
"type": "string",
"pattern": "^[0-9a-f]{64}$"
},
"parent_contract_id": {
"type": "string",
"pattern": "^[0-9a-f]{64}$"
},
"created_at": {
"type": "string",
"format": "date-time"
},
"title": {
"type": "string",
"maxLength": 200
},
"outcome": {
"type": "object",
"additionalProperties": false,
"required": [
"statement",
"success_definition",
"out_of_scope"
],
"properties": {
"statement": {
"type": "string",
"minLength": 1
},
"success_definition": {
"type": "string",
"minLength": 1
},
"out_of_scope": {
"type": "array",
"items": {
"type": "string",
"minLength": 1
}
}
}
},
"acceptance": {
"type": "object",
"additionalProperties": false,
"required": [
"criteria"
],
"properties": {
"criteria": {
"type": "array",
"minItems": 1,
"items": {
"type": "object",
"additionalProperties": false,
"required": [
"id",
"statement",
"observable_check",
"accepted_evidence"
],
"properties": {
"id": {
"type": "string",
"pattern": "^AC-[a-z0-9]+(-[a-z0-9]+)*$"
},
"statement": {
"type": "string",
"minLength": 1
},
"observable_check": {
"type": "string",
"minLength": 1,
"description": "An observable check: behavior, interface, or stored state - never a narration claim."
},
"accepted_evidence": {
"type": "array",
"minItems": 1,
"items": {
"enum": [
"observed",
"imported",
"asserted"
]
},
"uniqueItems": true,
"description": "Which section-4.5 trust classes may close this criterion. A criterion that does not accept 'asserted' never inherits completion from migration (A7/D3-5)."
},
"control": {
"type": "object",
"additionalProperties": false,
"required": [
"kind",
"rationale"
],
"properties": {
"kind": {
"enum": [
"regression",
"discrimination",
"exempt"
]
},
"rationale": {
"type": "string",
"minLength": 1
}
},
"description": "The negative control decided at contract authoring (section 6); 'exempt' is for minor prose/cosmetic criteria and says why."
}
}
}
}
}
},
"invariants": {
"type": "array",
"items": {
"type": "object",
"additionalProperties": false,
"required": [
"id",
"statement"
],
"properties": {
"id": {
"type": "string",
"pattern": "^INV-[a-z0-9]+(-[a-z0-9]+)*$"
},
"statement": {
"type": "string",
"minLength": 1
}
}
},
"description": "Global conditions that must hold at every boundary; violation is a stop, not an adaptation."
},
"scope": {
"type": "object",
"additionalProperties": false,
"required": [
"allowed_paths",
"allowed_command_classes",
"forbidden_operations"
],
"properties": {
"allowed_paths": {
"type": "array",
"items": {
"type": "string",
"minLength": 1
}
},
"allowed_command_classes": {
"type": "array",
"minItems": 1,
"items": {
"type": "string",
"minLength": 1
},
"uniqueItems": true,
"description": "Decidable only for declared gate commands the core validates and executes (section 8); an empty list is undeclared, not unrestricted; arbitrary shell is detection-plus-reporting, never prevention."
},
"forbidden_operations": {
"type": "array",
"minItems": 1,
"items": {
"type": "string",
"minLength": 1
},
"description": "Destructive and outward-facing operations, named explicitly."
}
}
},
"authorization": {
"type": "object",
"additionalProperties": false,
"required": [
"mechanism",
"authority",
"timestamp",
"boundaries",
"consent_checkpoints"
],
"properties": {
"mechanism": {
"enum": [
"plan_authorship",
"pre_authorization"
],
"description": "The binding mechanism ONLY (D3-1). plan_authorship = interactive git plans (the reviewed plan markdown is the consented artifact; authority = the session user). pre_authorization = unattended/non-git plans AND migration (the recorded pre-authorization is the migration request, D3-2). No third value exists. The citing record is the materialization-time journal 'approval' event - a record inside the contract cannot cite the contract's own content-addressed contract_id."
},
"authority": {
"type": "string",
"minLength": 1,
"description": "Who or what approved - an identity, never a model's confidence."
},
"timestamp": {
"type": "string",
"format": "date-time"
},
"boundaries": {
"type": "string",
"minLength": 1,
"description": "Boundaries of the pre-approval: what it does and does not cover."
},
"consent_checkpoints": {
"type": "array",
"items": {
"type": "string",
"minLength": 1
},
"description": "Carried verbatim from the plan; never paraphrased."
}
}
},
"permissions": {
"type": "object",
"additionalProperties": false,
"required": [
"granted",
"not_granted"
],
"properties": {
"granted": {
"type": "array",
"items": {
"$ref": "#/$defs/capability"
},
"uniqueItems": true
},
"not_granted": {
"type": "array",
"items": {
"$ref": "#/$defs/capability"
},
"uniqueItems": true,
"description": "Explicitly those NOT granted - the negative list is content, not filler."
}
}
},
"dependencies": {
"type": "array",
"items": {
"type": "object",
"additionalProperties": false,
"required": [
"name",
"kind",
"detail"
],
"properties": {
"name": {
"type": "string",
"minLength": 1
},
"kind": {
"enum": [
"external_system",
"credential",
"pinned_input"
]
},
"detail": {
"type": "string",
"minLength": 1,
"description": "For credentials: names only, never secrets."
}
}
}
},
"resource_envelope": {
"type": "object",
"additionalProperties": false,
"required": [
"limits"
],
"properties": {
"limits": {
"type": "array",
"minItems": 1,
"items": {
"type": "object",
"additionalProperties": false,
"required": [
"id",
"limit",
"unit",
"enforcement"
],
"properties": {
"id": {
"type": "string",
"pattern": "^[a-z][a-z0-9_]*$"
},
"limit": {
"type": "number",
"minimum": 0,
"description": "Malformed values (negative, non-numeric) are invalid; the semantic validator rejects them with the offending id named."
},
"unit": {
"type": "string",
"minLength": 1
},
"enforcement": {
"enum": [
"enforced",
"advisory"
],
"description": "enforced = a host adapter can actually stop the agent; advisory = instructed and reported only. Enforcement parity across hosts is never claimed (section 8)."
},
"metering_source": {
"type": "string",
"minLength": 1,
"description": "REQUIRED iff enforcement=enforced (A5): a host adapter reading real provider meters. authorize() never enforces on asserted meter data - an asserted meter degrades the limit to advisory and the record says so."
},
"reserve": {
"type": "number",
"minimum": 0,
"description": "OPTIONAL (section 8): budget held back from dispatch for verification, retry and resume. 0 <= reserve <= limit (the semantic validator enforces the ceiling and names the limit id); the dispatchable ceiling is limit - reserve."
}
},
"allOf": [
{
"if": {
"properties": {
"enforcement": {
"const": "enforced"
}
}
},
"then": {
"required": [
"metering_source"
],
"description": "A5: an enforced limit names its metering source."
}
}
]
}
}
}
},
"scheduling": {
"type": "object",
"additionalProperties": false,
"description": "Deterministic-authorization policy (RFC section 5), declared in the contract. Every field is optional: an absent field means the authorization core applies its shipped finite default (never unlimited); a plan may declare tighter or looser bounds.",
"properties": {
"starvation_threshold_events": {
"type": "integer",
"minimum": 1,
"description": "Journal-event count a ready-but-unselected task waits before the oldest-blocked-wait priority boost fires (A11; the clock is the journal, never the wall)."
},
"max_adaptations_per_task": {
"type": "integer",
"minimum": 0,
"description": "Maximum authorized adaptations whose affected_tasks include one task; bounded adaptation, loop termination."
},
"max_retries_per_gate": {
"type": "integer",
"minimum": 0,
"description": "Maximum authorized retry adaptations per acceptance criterion; blind retries are refused independently of this cap."
},
"handoff": {
"type": "object",
"additionalProperties": false,
"description": "Explicit handoff conditions (RFC section 5): when a fresh context or a cross-host resume is required.",
"properties": {
"fresh_context": {
"type": "string",
"minLength": 1,
"description": "Condition under which the session must be replaced by a fresh context."
},
"cross_host_resume": {
"type": "string",
"minLength": 1,
"description": "Condition under which the plan must resume on another host."
}
}
}
}
},
"tasks": {
"type": "array",
"minItems": 1,
"items": {
"type": "object",
"additionalProperties": false,
"required": [
"id",
"title",
"prerequisites",
"touched_surface",
"gate_intent"
],
"properties": {
"id": {
"type": "string",
"pattern": "^T-[a-z0-9]+(-[a-z0-9]+)*$"
},
"title": {
"type": "string",
"minLength": 1
},
"prerequisites": {
"type": "array",
"items": {
"type": "string",
"pattern": "^T-[a-z0-9]+(-[a-z0-9]+)*$"
},
"uniqueItems": true,
"description": "Stable task IDs whose outcomes must hold first. The semantic validator rejects dangling references and prerequisite cycles."
},
"touched_surface": {
"type": "array",
"items": {
"type": "string",
"minLength": 1
}
},
"gate_intent": {
"type": "array",
"items": {
"type": "object",
"additionalProperties": false,
"required": [
"criterion",
"check"
],
"properties": {
"criterion": {
"type": "string",
"pattern": "^AC-[a-z0-9]+(-[a-z0-9]+)*$"
},
"check": {
"type": "string",
"minLength": 1
}
}
},
"description": "Selected gate intent: which acceptance criterion each declared gate serves."
},
"parallel_safe": {
"type": "boolean",
"description": "Set by create (never inferred at execution): the task can run in a delegate without interleaving with sibling work - its touched surface does not overlap an in-progress sibling and it declares a bounded output. Absent means false."
}
}
}
}
},
"$defs": {
"capability": {
"enum": [
"gate_command_exec",
"fs_write_plan_scope",
"fs_write_repo_scope",
"git_operations",
"network_access",
"host_adapter_metering",
"agent_delegation",
"context_export",
"model_routing"
],
"description": "Closed capability set (section 3.2 Permissions). gate_command_exec: run DECLARED gate commands through the core's runner. fs_write_plan_scope / fs_write_repo_scope: writes inside the plan folder / inside allowed repo paths. git_operations: commits and fingerprinting. network_access: outbound fetch to named dependencies. host_adapter_metering: read real provider meters through an adapter. agent_delegation: sibling-plan fan-out under section 9's declared file ownership. context_export: the section-4.3 durability export command. model_routing: switch model tiers mid-plan (section 8) - an authority GRANT only; switching additionally requires the host ability of the same name, so the default posture stays fixed-model. Names outside this set are unsupported capabilities and are refused."
}
}
}
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness. | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is available | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |