Skills for Real Engineers. Straight from my .agents directory.
anomaly-alert
Agent skill fromhoangsonww/Claude-Code-Agent-Monitor
Identify anomalous sessions using Agent Monitor data — cost outliers from the pricing engine, token anomalies (cache miss spikes, compaction baseline surges), unusual event type ratios (PreToolUse/PostToolUse gaps, APIError clusters), behavioral deviations from workflow intelligence (complexity score outliers, error propagation anomalies), and sessions with abnormal metadata (extreme turn_count, high thinking_blocks, zero turn_duration).
How to install anomaly-alert
npx skills add hoangsonww/Claude-Code-Agent-Monitor# or add only this skill to Claude Code: copy the folder
plugins/ccam-insights/skills/anomaly-alert/
# into
~/.claude/skills/anomaly-alert/ (all projects)
.claude/skills/anomaly-alert/ (one project)
1# Anomaly Alert23Detect anomalous sessions in Claude Code Agent Monitor data.45## Input67The user provides: **$ARGUMENTS**89This may be:10- "all" or empty (default: check all anomaly types)11- "cost" for cost anomalies only12- "duration" for duration anomalies only13- "errors" for error rate anomalies only14- A sensitivity level: "strict" (1σ), "normal" (2σ), "relaxed" (3σ)1516## Procedure17181. **Fetch baseline data** from `http://localhost:4820`:19 - `GET /api/sessions?limit=500` — historical sessions for baseline20 - `GET /api/analytics` — aggregated metrics21 - `GET /api/pricing/cost` — cost data per session22232. **Compute baselines** for each metric:24 - Mean, median, standard deviation25 - P25, P75, P90, P95, P99 percentiles26 - Interquartile range (IQR) for robust outlier detection27283. **Detect anomalies** using statistical thresholds:2930 ### Cost Anomalies31 - Sessions costing >2σ above mean32 - Single sessions exceeding daily average33 - Sudden cost spikes (session-over-session increase >200%)3435 ### Duration Anomalies36 - Sessions lasting >2σ above mean duration37 - Extremely short sessions (<1 minute) that still incur cost38 - Sessions with unusual active-vs-idle ratios3940 ### Error Rate Anomalies41 - Sessions with error rates >2σ above baseline42 - New error types not seen in previous sessions43 - Sessions with >3 consecutive tool failures4445 ### Behavioral Anomalies46 - Unusual tool combinations not seen before47…
Excerpt shown under the repository's MIT license.
More skills in Claude-Code-Agent-Monitor
Similar skill repositories
An agentic skills framework & software development methodology that works.
A single CLAUDE.md file to improve Claude Code behavior, derived from Andrej Karpathy's observations on LLM coding pitfalls.
Public repository for Agent Skills
Makes your AI agent think like the laziest senior dev in the room. The best code is the code you never wrote.
Persistent Context Across Sessions for Every Agent – Captures everything your agent does during sessions, compresses it with AI, and injects relevant context back into future sessions. Works with Claude Code, OpenClaw, Codex, Gemini, Hermes, Copilot, OpenCode + More
