Higress is a cloud-native API gateway built on Istio and Envoy that routes and manages APIs and can be extended with WebAssembly plugins. It is used to manage APIs for AI models and MCP servers as well as general services, and the catalogue entries provide agent workflows and integrations for operating it.
Borrowing it
Nothing to install: this file belongs to higress-group/higress. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/higress-group/higress/main/.agents/skills/agent-session-monitor/SKILL.mdgit clone --depth 1 https://github.com/higress-group/higressWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/higress-group/higress/agent-session-monitor)<a href="https://agentmods.dev/skills/higress-group/higress/agent-session-monitor"><img src="https://agentmods.dev/badge/skills/higress-group/higress/agent-session-monitor/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/higress-group/higress/agent-session-monitor"><img src="https://agentmods.dev/badge/skills/higress-group/higress/agent-session-monitor.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector warn
SkillSpector: 1 finding, up to high
These are SkillSpector’s own severities. On a checked sample its high-severity flags on skills were ~96% false positives — a documented command, a public API, a “never do X” rule — so we show them as a caution to read, not a verdict. Why →
- high Data Exfiltration · line 366 Code or instructions that leak agent conversation context to external services, potentially exposing sensitive user interactions.Fix: Remove any code that sends prompts, responses, or session data externally. Preserve user privacy; never exfiltrate conversation content.
What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00053 | $0.03229 |
| Opus 5 | $0.00026 | $0.01614 |
| Sonnet 5 | $0.00011 | $0.00646 |
| Haiku 4.5 | $0.00005 | $0.00323 |
Grade A, and why
agent-session-monitor scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 12d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 377 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Overview
Real-time monitoring of Higress access logs, extracting ai_log JSON, grouping multi-turn conversations by session_id, and calculating token costs with visualization.
Core Features
- Real-time Log Monitoring: Monitors Higress access log files, parses new ai_log entries in real-time
- Log Rotation Support: Full logrotate support, automatically tracks access.log.1~5 etc.
- Incremental Parsing: Inode-based tracking, processes only new content, no duplicates
- Session Grouping: Associates multi-turn conversations by session_id (each turn is a separate request)
- Complete Conversation Tracking: Records messages, question, answer, reasoning, tool_calls for each turn
- Token Usage Tracking: Distinguishes input/output/reasoning/cached tokens
- Web Visualization: Browser-based UI with overview and session drill-down
- Real-time URL Generation: Clawdbot can generate observation links based on current session ID
- Background Processing: Independent process, continuously parses access logs
- State Persistence: Maintains parsing progress and session data across runs
Usage
1. Background Monitoring (Continuous)
# Parse Higress access logs (with log rotation support)
python3 main.py --log-path /var/log/proxy/access.log --output-dir ./sessions
# Filter by session key
python3 main.py --log-path /var/log/proxy/access.log --session-key <session-id>
# Scheduled task (incremental parsing every minute)
* * * * * python3 /path/to/main.py --log-path /var/log/proxy/access.log --output-dir /var/lib/sessions
2. Start Web UI (Recommended)
# Start web server
python3 scripts/webserver.py --data-dir ./sessions --port 8888
# Access in browser
open http://localhost:8888
Web UI features:
- 📊 Overview: View all session statistics and group by model
- 🔍 Session Details: Click session ID to drill down into complete conversation history
- 💬 Conversation Log: Display messages, question, answer, reasoning, tool_calls for each turn
- 💰 Cost Statistics: Real-time token usage and cost calculation
- 🔄 Auto Refresh: Updates every 30 seconds
What ships with it
11 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
- example/clawdbot_demo.py 2.7 KB runs code
- example/demo_v2.sh 2.6 KB runs code
- example/demo.sh 3.1 KB runs code
- example/test_access_v2.log 2.9 KB
- example/test_access.log 2.5 KB
- example/test_rotation.sh 4.6 KB runs code
- main.py 24 KB runs code
- QUICKSTART.md 4.1 KB
- README.md 1.9 KB
- scripts/cli.py 22 KB runs code
- scripts/webserver.py 28 KB runs code
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 12d ago First seen · 377 lines · 53 tokens per session scan A d386a470a394
agent-session-monitor is a skill published in the GitHub repository higress-group/higress (9,365 stars, last pushed yesterday), licensed Apache-2.0. It adds 53 tokens to every session and 3,229 once invoked, about $0.0003 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
API Gateway Testing
API gateway testing skill covering rate limiting validation, request routing, authentication proxy testing, load balancing verification, circuit breaker testing, and gateway configuration validation for Kong, Envoy, and AWS API Gateway.
serverless-expert
Design and implement production-grade serverless applications with optimal performance, cost efficiency, and scalability. Use when the user mentions serverless or FaaS, AWS Lambda, Azure Functions, Cloud Functions, cold starts, event-driven architecture, or API Gateway-fronted workloads.
kimss-mcp-setup
Configure the Kimss MCP server in Cursor, Windsurf, Claude Desktop, or other MCP clients using uvx, environment variables, and the seven v1 tools.
kimss-sdk
Integrate Kimss via dual-listener gateway (OpenAI /v1 or Anthropic /v1/messages + Agent-Id headers). Optional kimss package for control-plane register/usage only.
cli-eval
Create and run evaluation suites, watch live benchmark progress, view scorecards, compare model performance, and integrate eval runs with CI workflows from the CLI.
omni-mcp
Connect to the OmniRoute MCP server (110 tools, 3 transports: SSE/stdio/HTTP). Covers routing, cache, compression, memory, skills, providers, and audit tools across 33 permission scopes.