Layer 1 — agent framework laws · 235 documents
| ID | Document | Source |
|---|---|---|
| agent_invariants__08d0d1861c9bb388 | Complex tasks must be decomposed into atomic sub-tasks with explicit dependency graphs. Planning must include rollback strategies and replanning capability. | — |
| agent_invariants__55bef1f457793086 | Destructive operations require explicit human confirmation or pre-approved allow-lists with audit logging. | — |
| agent_invariants__2c656b98ac2ffe6d | Agent state must be serializable and checkpointed at every decision boundary to enable resumption after interruption and crash recovery. | — |
| agent_invariants__1522555553c6d09f | Agents must monitor token usage, implement truncation or summarization when approaching limits, and maintain coherence across chunks. | — |
| agent_invariants__e57b5c40c6286a25 | Agents must define tools with explicit input schemas, validate inputs before execution, handle timeouts, and degrade gracefully on failure. | — |
| agent_invariants__6d88dceff8b4ad2d | Agents require hierarchical memory: working memory for active context, short-term with relevance decay, and long-term with vector-based retrieval. | — |
| agent_invariants__8845cf8ae19ed365 | Complex tasks must be decomposed into atomic sub-tasks with explicit dependency graphs. Planning must include rollback strategies and replanning capability. | — |
| agent_invariants__0c53d0250072a954 | Destructive operations require explicit human confirmation or pre-approved allow-lists with audit logging. | — |
| agent_invariants__0ca85132ea125690 | External interactions must implement retry with exponential backoff, circuit breakers for cascading failure prevention, and graceful degradation. | — |
| agent_invariants__c09e8ac8e2b61be3 | Agent state must be serializable and checkpointed at every decision boundary to enable resumption after interruption and crash recovery. | — |
| agent_invariants__84d18a63afb7bc6c | Inter-agent communication must use structured message protocols with schema validation, streaming support, and event-driven callbacks. | — |
| agent_invariants__fa4d2e3ef9391505 | Agent access requires credential isolation, secure vaults, automatic key rotation, and least-privilege RBAC. | — |
| agent_invariants__277dbaae38024520 | External API calls must respect rate limits through token bucket or sliding window algorithms with automatic backoff. | — |
| agent_invariants__3510354ce2973082 | Agents must monitor token usage, implement truncation or summarization when approaching limits, and maintain coherence across chunks. | — |
| agent_invariants__3a92dfd2da889af2 | Inter-agent communication must use structured message protocols with schema validation, streaming support, and event-driven callbacks. | — |
| agent_invariants__edfa8dd855b46f6f | Agents must define tools with explicit input schemas, validate inputs before execution, handle timeouts, and degrade gracefully on failure. | — |
| agent_invariants__a6d7849dbb2bb9e0 | Agents require hierarchical memory: working memory for active context, short-term with relevance decay, and long-term with vector-based retrieval. | — |
| agent_invariants__7e62b994ad622f98 | Complex tasks must be decomposed into atomic sub-tasks with explicit dependency graphs. Planning must include rollback strategies and replanning capability. | — |
| agent_invariants__94f80e3256da7a18 | Destructive operations require explicit human confirmation or pre-approved allow-lists with audit logging. | — |
| agent_invariants__bdd212635ae1954a | External interactions must implement retry with exponential backoff, circuit breakers for cascading failure prevention, and graceful degradation. | — |
| agent_invariants__806b525bb3e6c74b | Agent state must be serializable and checkpointed at every decision boundary to enable resumption after interruption and crash recovery. | — |
| agent_invariants__58705558c8300483 | Inter-agent communication must use structured message protocols with schema validation, streaming support, and event-driven callbacks. | — |
| agent_invariants__bac435e8560ef5a4 | Agent access requires credential isolation, secure vaults, automatic key rotation, and least-privilege RBAC. | — |
| agent_invariants__7dd061a3acb4b9c6 | External API calls must respect rate limits through token bucket or sliding window algorithms with automatic backoff. | — |
| agent_invariants__a7a2c8c234f88cba | Agents must monitor token usage, implement truncation or summarization when approaching limits, and maintain coherence across chunks. | — |
| agent_invariants__f25a0b4e6f18cbd6 | Agents require hierarchical memory: working memory for active context, short-term with relevance decay, and long-term with vector-based retrieval. | — |
| agent_invariants__2949c79c1f906642 | Inter-agent communication must use structured message protocols with schema validation, streaming support, and event-driven callbacks. | — |
| agent_invariants__efbbf1f88538d1fa | Agents must define tools with explicit input schemas, validate inputs before execution, handle timeouts, and degrade gracefully on failure. | — |
| agent_invariants__6a640db7a7aacb78 | Agents require hierarchical memory: working memory for active context, short-term with relevance decay, and long-term with vector-based retrieval. | — |
| agent_invariants__e1838663985bc243 | Destructive operations require explicit human confirmation or pre-approved allow-lists with audit logging. | — |
| agent_invariants__042f6bf5e058cd68 | External interactions must implement retry with exponential backoff, circuit breakers for cascading failure prevention, and graceful degradation. | — |
| agent_invariants__d36562ce8bd01710 | Agent state must be serializable and checkpointed at every decision boundary to enable resumption after interruption and crash recovery. | — |
| agent_invariants__99423662d9594691 | Inter-agent communication must use structured message protocols with schema validation, streaming support, and event-driven callbacks. | — |
| agent_invariants__4c5910c57fcc8140 | Agent access requires credential isolation, secure vaults, automatic key rotation, and least-privilege RBAC. | — |
| agent_invariants__863c3c7c815e1cd4 | Agents require hierarchical memory: working memory for active context, short-term with relevance decay, and long-term with vector-based retrieval. | — |
| agent_invariants__f086ed7be12fbbf7 | Complex tasks must be decomposed into atomic sub-tasks with explicit dependency graphs. Planning must include rollback strategies and replanning capability. | — |
| agent_invariants__350e4900d56b4bac | External interactions must implement retry with exponential backoff, circuit breakers for cascading failure prevention, and graceful degradation. | — |
| agent_invariants__2065b1b5044eb8c2 | Agents must define tools with explicit input schemas, validate inputs before execution, handle timeouts, and degrade gracefully on failure. | — |
| agent_invariants__e7725e935496a81c | Agents require hierarchical memory: working memory for active context, short-term with relevance decay, and long-term with vector-based retrieval. | — |
| agent_invariants__65419ebc266c57c9 | Complex tasks must be decomposed into atomic sub-tasks with explicit dependency graphs. Planning must include rollback strategies and replanning capability. | — |
| agent_invariants__347ecca70340b3fd | Destructive operations require explicit human confirmation or pre-approved allow-lists with audit logging. | — |
| agent_invariants__10c42f46632f1acf | Agent state must be serializable and checkpointed at every decision boundary to enable resumption after interruption and crash recovery. | — |
| agent_invariants__d088056dcfd4cd88 | Inter-agent communication must use structured message protocols with schema validation, streaming support, and event-driven callbacks. | — |
| agent_invariants__04eb83c0f2d3d185 | Agents must monitor token usage, implement truncation or summarization when approaching limits, and maintain coherence across chunks. | — |
| agent_invariants__b5e2f5bb831cfb89 | Destructive operations require explicit human confirmation or pre-approved allow-lists with audit logging. | — |
| agent_invariants__a935db3cf53ae7dc | Inter-agent communication must use structured message protocols with schema validation, streaming support, and event-driven callbacks. | — |
| agent_invariants__0768b007383da7ef | Agents must define tools with explicit input schemas, validate inputs before execution, handle timeouts, and degrade gracefully on failure. | — |
| agent_invariants__771959710372b4df | Agents require hierarchical memory: working memory for active context, short-term with relevance decay, and long-term with vector-based retrieval. | — |
| agent_invariants__782a27729eab1c40 | Complex tasks must be decomposed into atomic sub-tasks with explicit dependency graphs. Planning must include rollback strategies and replanning capability. | — |
| agent_invariants__a1ad61f11560c346 | Destructive operations require explicit human confirmation or pre-approved allow-lists with audit logging. | — |