{"id":44616809,"url":"https://github.com/hernancapucci/agent-manifest","last_synced_at":"2026-03-12T18:53:09.522Z","repository":{"id":336646388,"uuid":"1149737461","full_name":"hernancapucci/agent-manifest","owner":"hernancapucci","description":"Declarative specification for AI agents to publish identity, capabilities and operational boundaries before interaction.","archived":false,"fork":false,"pushed_at":"2026-03-09T00:01:55.000Z","size":329,"stargazers_count":1,"open_issues_count":0,"forks_count":0,"subscribers_count":1,"default_branch":"main","last_synced_at":"2026-03-09T04:40:55.418Z","etag":null,"topics":["agent-boundaries","agent-governance","agent-manifest","agent-specification","ai-agents","ai-ethics","ai-safety","ai-specification","ai-standards","autonomous-systems","declarative-boundaries","json-schema","llm-agents","pre-execution-authority"],"latest_commit_sha":null,"homepage":"https://agent-manifest-spec.org/","language":"HTML","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"apache-2.0","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/hernancapucci.png","metadata":{"files":{"readme":"README.md","changelog":"CHANGELOG.md","contributing":"CONTRIBUTING.md","funding":null,"license":"LICENSE","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":"SECURITY.md","support":null,"governance":"GOVERNANCE.md","roadmap":"ROADMAP.md","authors":null,"dei":null,"publiccode":null,"codemeta":null,"zenodo":null,"notice":null,"maintainers":null,"copyright":null,"agents":null,"dco":null,"cla":null}},"created_at":"2026-02-04T13:15:50.000Z","updated_at":"2026-03-09T00:01:59.000Z","dependencies_parsed_at":"2026-03-02T01:12:35.392Z","dependency_job_id":null,"html_url":"https://github.com/hernancapucci/agent-manifest","commit_stats":null,"previous_names":["hernancapucci/agent-manifest"],"tags_count":3,"template":false,"template_full_name":null,"purl":"pkg:github/hernancapucci/agent-manifest","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/hernancapucci%2Fagent-manifest","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/hernancapucci%2Fagent-manifest/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/hernancapucci%2Fagent-manifest/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/hernancapucci%2Fagent-manifest/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/hernancapucci","download_url":"https://codeload.github.com/hernancapucci/agent-manifest/tar.gz/refs/heads/main","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/hernancapucci%2Fagent-manifest/sbom","scorecard":null,"host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":286080680,"owners_count":30438731,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2026-03-12T14:34:45.044Z","status":"ssl_error","status_checked_at":"2026-03-12T14:09:33.793Z","response_time":114,"last_error":"SSL_read: unexpected eof while reading","robots_txt_status":"success","robots_txt_updated_at":"2025-07-24T06:49:26.215Z","robots_txt_url":"https://github.com/robots.txt","online":false,"can_crawl_api":true,"host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":["agent-boundaries","agent-governance","agent-manifest","agent-specification","ai-agents","ai-ethics","ai-safety","ai-specification","ai-standards","autonomous-systems","declarative-boundaries","json-schema","llm-agents","pre-execution-authority"],"created_at":"2026-02-14T13:09:04.619Z","updated_at":"2026-03-12T18:53:09.517Z","avatar_url":"https://github.com/hernancapucci.png","language":"HTML","funding_links":[],"categories":[],"sub_categories":[],"readme":"[![DOI](https://zenodo.org/badge/DOI/10.5281/zenodo.18833956.svg)](https://doi.org/10.5281/zenodo.18833956)\n\n# Agent Manifest\n\n## Canonical Specification (v1.0 — Standards Track)\n\n- **HTML (rendered canonical document):**  \n  https://agent-manifest-spec.org/spec/v1.0/agent_manifest_v1.0.html\n\n- **Normative JSON Schema:**  \n  https://agent-manifest-spec.org/spec/v1.0/schema.json\n\n- **Markdown rendering (normatively identical):**  \n  [`spec/v1.0/spec.md`](./spec/v1.0/spec.md)\n\n  -----\n\n## Conceptual Architecture\n\nThe Agent Manifest ecosystem forms a layered declarative infrastructure for autonomous systems.\n\n![Agent Manifest Ecosystem Architecture](docs/architecture/agent-manifest-ecosystem.svg)\n\n-----\n  \nA minimal specification that requires autonomous systems to declare operational boundaries before execution.\n\nIt standardizes how AI agents declare identity, constraints, autonomy level, risk profile, and data handling prior to interacting with external systems.\n\nThis repository defines a Declaration Layer only.\n\nIt does not define behavior. It defines boundaries.\n\nAgent Manifest does not execute, validate, score, enforce, or decide.\n\n-----\n\n## 5-Minute Integration\n\nYou can integrate Agent Manifest in under five minutes.\n\n### 1. Create `manifest.json`\n\n```json\n{\n  \"$schema\": \"https://agent-manifest-spec.org/spec/v1.0/schema.json\",\n\n  \"manifest_version\": \"1.0\",\n  \"agent_id\": \"my.agent\",\n  \"agent_name\": \"My Agent\",\n  \"agent_version\": \"1.0.0\",\n\n  \"owner\": {\n    \"type\": \"organization\",\n    \"identifier\": \"My Company\"\n  },\n\n  \"purpose\": {\n    \"primary_code\": \"general.assistance\",\n    \"description\": \"Provides bounded assistance without irreversible execution.\"\n  },\n\n  \"forbidden_actions\": [\n    \"execute_irreversible_actions\"\n  ],\n\n  \"autonomy\": {\n    \"level\": 1\n  },\n\n  \"risk_profile\": {\n    \"level\": \"low\"\n  },\n\n  \"data_handling\": {\n    \"stores_personal_data\": false\n  },\n\n  \"stopping_authority\": {\n    \"stoppable_by\": [\"owner\"],\n    \"mechanism\": \"Manual disable via hosting interface.\"\n  },\n\n  \"audit_surface\": {\n    \"logging\": \"basic\",\n    \"reconstructability\": \"partial\"\n  },\n\n  \"contact\": {\n    \"email\": \"contact@mycompany.com\"\n  }\n}\n```\n\n### 2. Validate against the schema\n\n```bash\najv validate -s spec/v1.0/schema.json -d manifest.json\n```\n\n### 3. Commit it to your repository\n\nDone.\n\nYour agent now declares its boundary surface before execution.\n\n-----\n\n## Minimal Example\n\n```json\n{\n  \"$schema\": \"https://agent-manifest-spec.org/spec/v1.0/schema.json\",\n\n  \"manifest_version\": \"1.0\",\n  \"agent_id\": \"example.minimal.agent\",\n  \"agent_name\": \"Minimal Example Agent\",\n  \"agent_version\": \"1.0.0\",\n\n  \"owner\": {\n    \"type\": \"individual\",\n    \"identifier\": \"Example Owner\"\n  },\n\n  \"purpose\": {\n    \"primary_code\": \"general.assistance\",\n    \"description\": \"Provides bounded general assistance without executing irreversible actions.\"\n  },\n\n  \"forbidden_actions\": [\n    \"execute_financial_transactions\",\n    \"access_private_data_without_consent\"\n  ],\n\n  \"autonomy\": {\n    \"level\": 1\n  },\n\n  \"risk_profile\": {\n    \"level\": \"low\"\n  },\n\n  \"data_handling\": {\n    \"stores_personal_data\": false\n  },\n\n  \"stopping_authority\": {\n    \"stoppable_by\": [\"owner\"],\n    \"mechanism\": \"Agent can be halted via manual override by the declared owner.\",\n    \"stages\": [\n      \"pre-execution\",\n      \"mid-execution\"\n    ]\n  },\n\n  \"audit_surface\": {\n    \"logging\": \"basic\",\n    \"reconstructability\": \"partial\",\n    \"opacity_declared\": false\n  },\n\n  \"contact\": {\n    \"email\": \"contact@example.com\"\n  }\n}\n```\n\nThis example is non-normative and provided for structural clarity.  \nFull structural reference: [`spec/v1.0/spec.md`](./spec/v1.0/spec.md)\n\nFor additional examples see:  \n[`examples/`](./examples/)\n\n-----\n\n## Why Agent Manifest exists\n\nAs AI systems become increasingly autonomous, interactions between agents, humans, APIs, and infrastructures require clarity prior to execution.\n\nMost systems describe capabilities.  \nFew describe constraints.\n\nAgent Manifest introduces a structural principle:\n\n\u003e Autonomous systems should declare boundaries before action.\n\nThis specification provides a public declaration layer — not a runtime, not a framework, not a governance engine.\n\nIt is designed to be:\n\n- Minimal  \n- Composable  \n- Execution-agnostic  \n- Forward-compatible  \n\n-----\n\n## What Agent Manifest is\n\nAgent Manifest is a declarative surface that may include:\n\n- Agent identity  \n- Operational domain (Scope)  \n- Autonomy level  \n- Tool access scope  \n- Data handling guarantees  \n- Risk profile  \n- Human oversight conditions  \n- Version commitments  \n\nIt is intentionally minimal and does not mandate internal architecture.\n\n-----\n\n## What Agent Manifest is not\n\nAgent Manifest does not:\n\n- Execute agents  \n- Enforce compliance  \n- Guarantee safety  \n- Replace governance frameworks  \n- Certify correctness  \n- Score risk  \n- Monitor behavior  \n\nIt defines what an agent *declares* — not what it *does*.\n\nValidation, scoring, auditing, and enforcement belong to separate systems.\n\n-----\n\n## Stability\n\nThe normative specification contract is defined in:\n\n- [`spec/v1.0/spec.md`](./spec/v1.0/spec.md)  \n- [`spec/v1.0/schema.json`](./spec/v1.0/schema.json)  \n\nThe constitutional core is defined in:\n\n- [`CORE_PRINCIPLES.md`](./CORE_PRINCIPLES.md)\n\nSee: [`STABILITY.md`](./STABILITY.md) and [`VERSIONING_POLICY.md`](./VERSIONING_POLICY.md)\n\n-----\n\n## Start here\n\nIf you are new to Agent Manifest, read in this order:\n\n1. [WHY_THIS_EXISTS.md](./WHY_THIS_EXISTS.md)  \n2. [WHAT_THIS_IS.md](./WHAT_THIS_IS.md)  \n3. [CORE_PRINCIPLES.md](./CORE_PRINCIPLES.md)  \n4. [`spec/v1.0/`](./spec/v1.0/)  \n5. [`examples/`](./examples/)  \n\n-----\n\n## Design Rationale\n\nThe structural reasoning behind the specification is documented in:\n\n[DESIGN_RATIONALE.md](./DESIGN_RATIONALE.md)\n\nThis document explains the architectural decisions, boundary-first philosophy, autonomy model, and execution-agnostic design principles that shape Agent Manifest.\n\n-----\n\n## Security \u0026 Governance Context\n\nAgent Manifest may be referenced during security and governance review of autonomous systems.\n\nFor structured review guidance, see:\n\n[SECURITY_REVIEW_CONTEXT.md](./SECURITY_REVIEW_CONTEXT.md)\n\n-----\n\n## Architectural Layering\n\nAgent Manifest operates strictly at the Declaration Layer.\n\nIt does not execute agents.  \nIt does not enforce policy.  \nIt does not perform runtime validation.\n\nInstead, it standardizes how agents declare:\n\n- Identity  \n- Purpose  \n- Boundaries  \n- Autonomy level  \n- Risk profile  \n- Data handling commitments  \n\nThe ecosystem separates into three distinct layers:\n\n### 1. Declaration Layer\n\nAgent Manifest (this repository)  \nDefines how boundaries and commitments are declared.\n\n### 2. Enforcement Layer\n\nValidators, policy engines, audit systems, and compliance tooling.  \nThese systems verify whether declarations are internally consistent and externally respected.\n\n### 3. Execution Layer\n\nAgents and runtimes that perform actions.\n\nWithout structured declaration, validation is fragile.  \nWithout validation, accountability erodes.\n\n-----\n\n## Ecosystem Infrastructure\n\nAgent Manifest defines the declaration layer of a broader open ecosystem.\n\nOperational components such as the generator, dataset repository, registry, and discovery endpoint are documented in:\n\n[docs/ecosystem-map.md](docs/ecosystem-map.md)\n\nThese infrastructure components operate independently from the core specification and may evolve without modifying the normative spec.\n\n-----\n\n## Governance\n\nThis repository defines the open specification.\n\nInterpretations, validators, scoring engines, and audit systems may be built independently and are not part of this core specification.\n\nProposals for modification must preserve the core principles defined in  \n[CORE_PRINCIPLES.md](./CORE_PRINCIPLES.md).\n\n-----\n\n## Versioning\n\nAgent Manifest follows semantic versioning:\n\n- MAJOR: structural or constitutional changes  \n- MINOR: compatible clarifications  \n- PATCH: editorial or formatting updates  \n\nDocumentation and examples may evolve while `manifest_version` remains `1.0` until a formal spec bump is declared.\n\n-----\n\n## Long-term Intent\n\nAgent Manifest proposes a structural layer for autonomous systems:\n\nAgents should be able to declare who they are, what they can do, and where their boundaries lie before interaction begins.\n\nThis specification remains neutral, minimal, and open.\n\nIts value emerges through adoption.\n\n-----\n\n## Foundational Doctrine\n\nAgent Manifest is grounded in the principle of Pre-Execution Authority.\n\nBefore autonomy, execution, or capability, an agent must declare the authority framework under which it operates.\n\nSee:  \n[Pre-Execution Authority](./foundations/pre-execution-authority/README.md)\n\n----\n\n## Citation\n\nThe citable archived version of Agent Manifest v1.0 is available on Zenodo:\n\nCore Declarative Specification (Standard):  \nhttps://doi.org/10.5281/zenodo.18833956  \n\nCompanion analysis:  \nhttps://doi.org/10.5281/zenodo.18834845\n\n\n## Authorship\n\nAgent Manifest was conceived and authored by  \n**Hernán Alfredo Capucci** (2026)\n\nThe specification is maintained as an open standard.\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fhernancapucci%2Fagent-manifest","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fhernancapucci%2Fagent-manifest","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fhernancapucci%2Fagent-manifest/lists"}