docs(#6759): add normative-spec doc-update trigger to AGENTS.md

AGENTS.md

@@ -21,6 +21,7 @@ Fullsend is a platform for fully autonomous agentic development for Git-hosted o
 - When modifying CLI flag help text, defaults, or behavior in `internal/cli/`, update the corresponding reference page in `docs/cli/` and search `docs/guides/` for tables or descriptions that reference the same flag. CLI reference pages are manually maintained and will not auto-update.
 - When extending a Go interface with new methods, see [Go Code § Interface documentation](docs/contributing/go-code.md) for sync requirements.
 - When adding a new skill under `skills/`, check the user-facing guides for relevant cross-reference opportunities: `docs/guides/user/customizing-with-skills.md` (skill catalog and usage), `docs/guides/user/bring-your-own-agent.md` (agent composition and tuning), and `docs/guides/README.md` (guide index). Add a brief cross-reference or section pointer if the new skill fills a gap in those guides.
+- When adding, removing, or modifying fields in `event_payload` (`buildEventPayload` in `internal/harnessdispatch/project.go`) or normalized-event structures (`internal/normevent/`), update the projection table in the normative event specification at `docs/normative/normalized-event/v1/README.md` and the resolution description in `docs/contributing/harness-fields.md`.
 - **Per-org installation mode is deprecated** ([ADR 0044](docs/ADRs/0044-deprecate-per-org-installation-mode.md)) and is being removed. This applies to human contributors and agents alike: do not add or extend org-mode-specific content in docs or code, and when reviewing a PR that touches org-mode content, flag it as referencing deprecated functionality rather than engaging with the org/repo-mode distinction as active architecture. Per-repo is the sole supported installation model going forward.
 - When writing skills, documentation, or guides that reference the fullsend agent/skill/sub-agent inventory (agent names, skill names, harness configs, sub-agent rosters), use runtime discovery commands against `fullsend-ai/agents` rather than hardcoded tables or static listings. The agents repo evolves independently and hardcoded references go stale. See the `author-fullsend-augmentations` skill for discovery patterns.
 
@@ -39,6 +40,7 @@ Detailed guidance lives in `docs/contributing/` and topic-specific guides under
 | [Forge Abstraction](docs/contributing/forge-abstraction.md) | Adding forge operations — covers `forge.Client` interface rules |
 | [Harness Composition](docs/contributing/harness-composition.md) | Changing merge functions in `internal/harness/` — covers the invariant between compose and forge merge functions |
 | [Harness Field Reference](docs/contributing/harness-fields.md) | Adding or modifying fields in `Harness` or `ForgeConfig` — covers field classifications, merge rules, and the `ForgeConfig` struct |
+| [Normative Event Specification](docs/normative/normalized-event/v1/README.md) | Adding or modifying fields in `event_payload` (`buildEventPayload`), normalized-event structures under `internal/normevent/`, or adapters — covers the projection table, transition vocabulary, adapter contracts, and CEL trigger examples |
 | [CEL Triggers](docs/contributing/cel-triggers.md) | Writing or reviewing harness `trigger` CEL expressions or `.feature` CEL filters — covers normalized transition kinds |
 | [ADRs](docs/contributing/adrs.md) | Touching `docs/ADRs/` or reviewing ADR changes — covers immutability and status rules |
 | [Sandbox Topology](docs/contributing/sandbox-topology.md) | Modifying sandbox images, CI image pulling, or agent harness configs |