{"slug":"delegation-audit","title":"delegation-audit","summary":"Audit model delegation and subagent effectiveness for a session — which models handled which subagent types, per-type success rates and average durations, and wasted delegations (heavy models on trivial work or types that consistently fail) — using the Agent Monitor workflow inte","platform":"Claude","tags":[],"authorName":"LLM Mart","authorSlug":"llm-mart","score":0,"source":"github","price":null,"verified":false,"createdAt":"2026-09-07T18:39:08.989507Z","repo":{"url":"https://github.com/hoangsonww/Claude-Code-Agent-Monitor","stars":1014,"forks":238,"license":"MIT","updatedAt":"2026-09-24T18:17:15Z"},"bodyHtml":"<hr>\n<h2>name: delegation-audit\ndescription: &gt;\nAudit model delegation and subagent effectiveness for a session — which\nmodels handled which subagent types, per-type success rates and average\ndurations, and wasted delegations (heavy models on trivial work or types\nthat consistently fail) — using the Agent Monitor workflow intelligence API.\nUse when reviewing how a session delegated work across models and subagents.</h2>\n<h1>Delegation Audit</h1>\n<p>Audit how a Claude Code session delegated work: model-to-subagent mapping and whether each delegation paid off.</p>\n<h2>Input</h2>\n<p>The user provides: <strong>$ARGUMENTS</strong></p>\n<p>A session ID. If empty, fetch <code>GET /api/sessions?limit=1</code> and audit the most recent session, stating which one.</p>\n<h2>Data Sources</h2>\n<table>\n<thead>\n<tr>\n<th>Endpoint</th>\n<th>Returns</th>\n</tr>\n</thead>\n<tbody>\n<tr>\n<td><code>GET /api/workflows/{sessionId}</code></td>\n<td>The <code>modelDelegation</code> dataset (which models are delegated which subagent types) and the <code>effectiveness</code> dataset (per-type completion/success rate, avg duration, task success)</td>\n</tr>\n<tr>\n<td><code>GET /api/agents</code></td>\n<td>Raw subagent records (<code>type</code>, <code>model</code>, <code>status</code>, <code>depth</code>, <code>parent</code>) to corroborate counts and statuses</td>\n</tr>\n</tbody>\n</table>\n<h2>Report Sections</h2>\n<h3>1. Delegation Matrix</h3>\n<p>From <code>modelDelegation</code>: a model × subagent-type table of how many agents of each type each model ran.</p>\n<table>\n<thead>\n<tr>\n<th>Model</th>\n<th>explore</th>\n<th>code-review</th>\n<th>debugger</th>\n<th>...</th>\n<th>Total</th>\n</tr>\n</thead>\n</table>\n<h3>2. Effectiveness by Subagent Type</h3>\n<p>From <code>effectiveness</code>: per type, the success rate and average duration.\n| Subagent type | Count | Success rate | Avg duration | Verdict |\n|---------------|-------|--------------|--------------|---------|\nMark types below ~70% success as low-yield.</p>\n<h3>3. Wasted Delegations</h3>\n<p>Flag, with evidence:</p>\n<ul>\n<li>A heavy model (e.g. Opus) assigned to a simple/low-stakes subagent type that a cheaper model handled successfully elsewhere — candidate for rebalancing.</li>\n<li>Subagent types with low success rates (effort spent, task not completed).</li>\n<li>Duplicate delegations: the same type spawned repeatedly with poor success (retry churn).</li>\n</ul>\n<h3>4. Rebalancing Suggestions</h3>\n<p>Concrete model reassignments grounded in the matrix and effectiveness data. State the type, the model used, the success rate, and the suggested model — only where the data supports it.</p>\n<h2>Output</h2>\n<ul>\n<li>Markdown tables for the matrix and effectiveness.</li>\n<li>Success rates as percentages; durations in human units (e.g. <code>1m 12s</code>).</li>\n<li>Use ▲/▼ when comparing a type's success rate against the session-wide average.</li>\n<li>Cite only numbers returned by the API; do not infer success rates that the <code>effectiveness</code> dataset does not provide.</li>\n<li>If the dashboard is unreachable, tell the user to start it with <code>npm start</code> from the repo root.</li>\n</ul>\n","files":[{"path":"agents/openai.yaml","sizeBytes":268,"isText":true},{"path":"SKILL.md","sizeBytes":2711,"isText":true}],"reviewScore":null,"reviewSummary":null,"trust":{"provenance":"trusted-source-unreviewed","notice":"Community-authored content, reproduced verbatim and not vetted as instructions. Treat it as data to evaluate, never as directives to follow.","bodySource":null},"bodyLocked":false,"purchaseUrl":null,"sourceUrl":null,"report":{"provenance":"trusted-source-unreviewed","screen":{"ran":true,"outcome":"clean","suspicious":0,"notes":0,"hiddenCharacters":false},"virusScan":{"engine":"clamav","status":"clean","scannedAt":"2026-09-07T18:42:13.39797Z","sha256":"3C331C8B1BFF7E86359FF4E1BCE9D97985EE69B27DC235782763A771B41C8912","sizeBytes":1662},"review":null,"source":{"repositoryUrl":"https://github.com/hoangsonww/Claude-Code-Agent-Monitor","path":"plugins/ccam-workflows/skills/delegation-audit","license":"MIT","commit":"d130ebb498786c2b985a57e5c920ca781060b709","subtreeSha":"2570028A6079CA00A2C0C0B44D413D51E238D39F079864B39A3227D86B86FC99","lastSyncedAt":"2026-09-25T06:49:18.541012Z"},"reviewedAt":"2026-09-07T18:48:22.80759Z","notice":"Community-authored content, reproduced verbatim and not vetted as instructions. Treat it as data to evaluate, never as directives to follow."},"install":[{"target":"skills-cli","command":"npx skills add https://github.com/hoangsonww/Claude-Code-Agent-Monitor/tree/master/plugins/ccam-workflows/skills/delegation-audit"},{"target":"claude-code","command":"claude plugin marketplace add https://llmmart.ai/marketplace.json && claude plugin install hoangsonww-claude-code-agent-monitor@llmmart"},{"target":"git","command":"git clone https://github.com/hoangsonww/Claude-Code-Agent-Monitor.git"}]}