fix(uniswap-integrations): Opus 5 migration - real MCP tool names, bounded reports, no fabricated cost field

Broken references:
- refine-linear-task: linear_search_issues / linear_update_issue /
  linear_add_comment are not tools on mcp.linear.app. Replaced with the
  real surface: get_issue, list_issues, save_issue, save_comment.
- daily-standup: mcp__github__get_pull_request no longer exists on the
  GitHub MCP; it is pull_request_read.
- orchestrate-deployment: agent-orchestrator-agent lives in the
  development-codebase-tools plugin, so it needs the plugin-qualified
  dispatch name.

Invented numbers:
- deploy-guide output shape demanded a monthly cost estimate with no
  measured input. Made the field optional, required a named pricing
  source, and said plainly not to derive one from recall.

Output length (effort settings do not shorten authored deliverables):
- Bounded the datadog-cost-tracker and investigate-incident reports.

Delegation:
- Restated the deploy agent split as a ceiling and added a
  when-NOT-to-delegate rule.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

Author: wkoutre

CLAUDE.md

@@ -230,7 +230,7 @@ All plugins follow semantic versioning (semver). Key versioning rules:
 | development-productivity   | 2.4.1   |
 | skill-management           | 1.2.0   |
 | spec-workflow              | 2.1.0   |
-| uniswap-integrations       | 2.6.1   |
+| uniswap-integrations       | 2.7.0   |
 
 **Note:** Keep this table updated when versions change.
 

packages/plugins/uniswap-integrations/.claude-plugin/plugin.json

@@ -1,6 +1,6 @@
 {
   "name": "uniswap-integrations",
-  "version": "2.6.1",
+  "version": "2.7.0",
   "description": "External service integrations (Linear, Notion, Nx, Chrome DevTools, GitHub, Slack, Amplitude, Datadog) and deployment orchestration",
   "author": {
     "name": "Uniswap Labs",

packages/plugins/uniswap-integrations/skills/daily-standup/standup-guide.md

@@ -168,7 +168,7 @@ The workflow will:
   - Use `mcp__github__search_issues` with queries like:
     - `org:Uniswap author:<username> created:>YYYY-MM-DD` (PRs/issues created)
     - `org:Uniswap involves:<username> updated:>YYYY-MM-DD` (all activity including comments, reviews)
-  - Get PR details for any PRs found using `mcp__github__get_pull_request`
+  - Get PR details for any PRs found using `mcp__github__pull_request_read`
   - Include PR status (open/merged/closed) and review status
 - **Fetch Linear Issues**:
   - Call Linear MCP with the assignee parameter

packages/plugins/uniswap-integrations/skills/datadog-cost-tracker/SKILL.md

@@ -33,6 +33,10 @@ Analyze Datadog ingestion costs across logs, APM spans, and other estimated-usag
 
 ## Output Format
 
+Keep the whole report under ~400 words outside the tables. One or two sentences per section; the
+ranked table carries the detail. Do not restate the table in prose, and do not pad sections that
+have nothing to report — write "none" and move on.
+
 **Cost Summary** — total estimated ingestion for the period by signal type.
 
 **Top Ingesters** — ranked table: service | volume | WoW delta | notes.

packages/plugins/uniswap-integrations/skills/investigate-incident/SKILL.md

@@ -32,6 +32,9 @@ Investigate production incidents by querying Datadog for logs, metrics, and trac
 
 ## Output Format
 
+Keep the report under ~500 words outside quoted log lines and metric values. Evidence is what
+earns length; narration is not. Omit a section that has nothing to report rather than padding it.
+
 **Incident Summary**
 
 - What happened, when it started, and estimated user/system impact

packages/plugins/uniswap-integrations/skills/orchestrate-deployment/deploy-guide.md

@@ -1,8 +1,9 @@
 ---
 description: Orchestrate deployment pipelines, infrastructure setup, and CI/CD configuration using specialized deployment agents.
 argument-hint: <target> [--strategy blue-green|canary|rolling] [--environment dev|staging|prod] [--dry-run]
-allowed-tools: Read(*), Write(*), Task(subagent_type:cicd-agent), Task(subagent_type:infrastructure-agent), Task(subagent_type:agent-orchestrator-agent)
-# Note: agent-orchestrator-agent is from development-codebase-tools plugin (optional - see fallback below)
+allowed-tools: Read(*), Write(*), Task(subagent_type:cicd-agent), Task(subagent_type:infrastructure-agent), Task(subagent_type:development-codebase-tools:agent-orchestrator-agent)
+# Note: agent-orchestrator-agent is from the development-codebase-tools plugin, so it needs the
+# plugin-qualified dispatch name (optional - see fallback below)
 ---
 
 ## Inputs
@@ -36,7 +37,9 @@ Orchestrate complete deployment workflow through specialized agents:
 
 ## Orchestration Strategy
 
-> **Note**: This skill uses **agent-orchestrator-agent** from the development-codebase-tools plugin when available. If not installed, the skill will execute infrastructure-agent and cicd-agent sequentially instead of in parallel coordination.
+> **Note**: This skill uses **development-codebase-tools:agent-orchestrator-agent** when that plugin is installed. If not, execute infrastructure-agent and cicd-agent sequentially instead of in parallel coordination.
+
+**Delegate only when the work warrants it.** The agent split below is a ceiling, not a quota: at most one infrastructure-agent and one cicd-agent per deployment. When a phase finishes in a handful of direct tool calls (reading an existing workflow file, a single `kubectl rollout status`, a dry-run with no infrastructure changes), do it inline rather than spawning an agent for it.
 
 ### Phase 1: Pre-Deployment Analysis
 
@@ -179,8 +182,13 @@ Use **cicd-agent** for observability setup:
       loadBalancing: object;
       networking: object;
     };
-    costs: {
-      estimated: string; // Monthly cost estimate
+    // Omit `costs` entirely unless a real pricing source was queried (cloud pricing API,
+    // `infracost breakdown`, an existing billing dashboard). Do not derive a monthly figure
+    // from instance types and rate cards you recalled rather than read - a fabricated number
+    // reads as measured once it lands in the report.
+    costs?: {
+      estimated: string; // Monthly cost estimate, with the source that produced it
+      source: string; // e.g. "infracost breakdown", "AWS Pricing API"
       breakdown: object;
     };
   };

packages/plugins/uniswap-integrations/skills/refine-linear-task/SKILL.md

@@ -1,6 +1,6 @@
 ---
 description: Refine and enhance Linear task descriptions. Use when user says "refine this Linear task", "improve task description", "make this task clearer", "enhance Linear issue", or needs to improve clarity, completeness, and actionability of Linear issues.
-allowed-tools: Read(*), Glob(*), Grep(*), Task(subagent_type:Explore), WebSearch(*), WebFetch(*), mcp__linear__linear_search_issues(*), mcp__linear__linear_update_issue(*), mcp__linear__linear_add_comment(*)
+allowed-tools: Read(*), Glob(*), Grep(*), Task(subagent_type:Explore), WebSearch(*), WebFetch(*), mcp__linear__get_issue(*), mcp__linear__list_issues(*), mcp__linear__save_issue(*), mcp__linear__save_comment(*)
 ---
 
 # Linear Task Refiner
@@ -37,7 +37,7 @@ Parse from request:
 
 ### Phase 1: Fetch and Analyze Current Task
 
-1. **Fetch Issue Details** via Linear MCP tools
+1. **Fetch Issue Details** via `mcp__linear__get_issue` (or `mcp__linear__list_issues` when resolving an issue from a partial reference)
 2. **Display Current State**: Show issue details and description
 3. **Analyze for Gaps**: Evaluate against criteria:
    - Clarity: Is the problem/goal clearly stated?
@@ -75,8 +75,8 @@ Create comprehensive, well-structured description:
 
 ### Phase 5: Update Linear (with approval)
 
-1. **Update Issue Description** via Linear MCP
-2. **Add Comment** (optional): Note the refinement
+1. **Update Issue Description** via `mcp__linear__save_issue`
+2. **Add Comment** (optional) via `mcp__linear__save_comment`: note the refinement
 3. **Confirm Success**: Provide issue URL
 
 ## Focus Areas