From 0589794c56b40eb5bf1a349b318dca03b06adb44 Mon Sep 17 00:00:00 2001 From: bdchatham Date: Sun, 31 May 2026 12:26:05 -0700 Subject: [PATCH] docs: wire /brevity into Tide working agreements PR #89 landed the brevity skill with user-invocation routing. Layer 2: proactive use by referencing the skill in Tide's top-level working agreements so agents apply it without explicit user invocation. - CLAUDE.md: one bullet under Working Agreements, matching section cadence (25 words, imperative). - AGENTS.md: one paragraph under Working Agreement covering all 15 agents in the roster. Surface scoped to PR descriptions + in-code comments (matches the skill's MVP); deferred surfaces (doc/design, runbooks, memory writes, mid-conversation chat) enumerated explicitly with a pointer to references/guardrails.md. Skip language inverted on coral feedback: "The skill self-determines floor; agents do not pre-skip." Removes the bypassable threshold. Coral cross-review (product-engineer + reviewer): two HOLD findings (design-notes scope expansion + bullet length) resolved; cleared. Follow-ups tracked: - #90: portable-scrub + sync to user-scope (after 2 real-use cycles) - #91: persona-line /brevity reference in each .claude/agents/ file for sub-agent dispatch coverage Co-Authored-By: Claude Opus 4.7 --- AGENTS.md | 2 ++ CLAUDE.md | 1 + 2 files changed, 3 insertions(+) diff --git a/AGENTS.md b/AGENTS.md index ee5a1f81..b516edc7 100644 --- a/AGENTS.md +++ b/AGENTS.md @@ -32,6 +32,8 @@ All design work follows the constitution at `design/constitution/constitution.md 4. **Errors Are Interface** — every error condition is documented 5. **Tests Prove Interfaces** — concrete test specs for every interface +**Output discipline.** Every agent in the roster authors PR descriptions and in-code comments. Before any agent ships a PR body or writes a WHY-style comment, it applies `/brevity` (`.claude/skills/brevity/`). The skill's 8 rules + 5-row rationalization table hold against the verbose-by-default biases an LLM agent produces under pressure. The skill self-determines when input is at floor — agents do not pre-skip. Doc/design, runbooks, memory writes, and mid-conversation chat are out of scope for the skill today; see its `references/guardrails.md` for un-defer triggers. + ## Cross-Component Interfaces The most critical contracts between components. The **provider owns** the interface; consumers adapt. diff --git a/CLAUDE.md b/CLAUDE.md index e7ff7d2d..84b8e30a 100644 --- a/CLAUDE.md +++ b/CLAUDE.md @@ -25,6 +25,7 @@ The single source of truth for all cross-component interfaces is `tide/interface - **Completion signaling:** `/dev/termination-log` is primary (survives pod termination). `status.json` is advisory only. - **Exit codes:** Granular codes (10-52) from runtimes. Operator groups them for retry/fail decisions. - **Commit style:** Conventional commits (`feat:`, `fix:`, `docs:`, `refactor:`). Reference the component in scope. +- **Brevity discipline:** Apply `/brevity` (`.claude/skills/brevity/`) before writing PR bodies or in-code comments. The skill self-determines floor; agents do not pre-skip. ## Key File Locations - `design/constitution/constitution.md` — governing document for all design work