Local log scanning and secret redaction
Finds supported local agent logs, then parses metrics and redacts secrets on your machine before any upload step.
Agent Analyzer
Agent Analyzer is a local-first profiler for agentic coding logs. It scans supported logs on your machine, redacts secrets locally, and turns a sanitized report into a personalized plugin or setup pack.
Local profiler for agentic coding log waste
Agent Analyzer is a local-first profiler for agentic coding logs. It scans supported logs on your machine, redacts secrets locally, and produces a sanitized report that can be turned into a personalized plugin or setup pack.
The product is aimed at people who use tools such as Claude Code, Codex, GitHub Copilot, OpenCode, Cursor, Kiro, Antigravity, or Claude Desktop and want to understand where sessions waste tokens. Its core job is to identify avoidable context bloat, rereads, retry loops, and noisy tool output, then convert those findings into concrete guidance for future sessions.
What Agent Analyzer does
Deterministic report generation
Uses a deterministic parser and server-side rendering instead of an LLM reading the raw logs, which keeps the analysis approach predictable.
Token-waste detection
Detects waste patterns such as repeated reads, oversized shell output, retry loops, broad discovery, and context-growth spikes.
Custom plugin output
Turns the sanitized findings into a personalized plugin with targeted skills, slash commands, setup guidance, and vetted tool recommendations.
Reviewable report pack
Shows a report preview with estimated token volume, session timeline, and top problems so users can review the waste before changing anything.
Privacy-first upload boundary
Asks before upload and only sends the sanitized report JSON, with raw transcripts and raw SQLite stores kept local in the documented flow.
Common ways to use Agent Analyzer
Review personal agent workflow waste
Use it after a few coding sessions to see whether rereads, retry loops, or large shell outputs are inflating token use.
Shape a repeatable team workflow
Run it before standardizing a team setup so you can convert real log patterns into narrower commands, skills, and AGENTS.md guidance.
Audit session patterns before making changes
Use the generated report pack to understand where context growth or noisy tools are appearing before changing your setup.
Keep sensitive logs local
Apply it when you want local-first analysis and do not want raw transcripts or logs sent through the public web flow.
Validate workflow changes with evidence
Use the benchmark-backed output when comparing a candidate reducer or workflow change against measured evidence from repeated runs.
Pros and Cons
Pros
- Keeps raw agent logs on the user’s machine during parsing and redaction.
- Uses a deterministic analysis flow with a reviewable sanitized report before upload.
- Produces concrete outputs such as report packs, custom skills, slash commands, and AGENTS.md-style guidance.
- Backs recommendations with repeated benchmark methodology rather than only a single run or generic advice.
Cons
- The product is intentionally conservative and does not promise savings for every token category or workflow.
- The pricing page was not accessible in the provided evidence, so plan details and limits are unclear from the source.
- The generated guidance is based on supported logs and benchmarked patterns, which may be narrower than a general-purpose optimization tool.
FAQ
How does Agent Analyzer handle my logs?
Agent Analyzer runs a read-only local scan on supported agent logs. It parses and redacts data on your machine, then asks before uploading only the sanitized report JSON needed to build the plugin.
What kinds of inputs does it analyze?
The public flow is designed for supported local agent logs and bounded SQLite rows. The security page says it finds logs locally, redacts secrets locally, and does not receive raw transcripts through the web upload path.
What do I get after the scan?
The output is a report pack and a custom plugin or skill set tailored to the token-leak patterns found in your logs. The site also describes copy-ready AGENTS.md lines and vetted tool recommendations as part of the generated guidance.
What limitations should I expect?
The benchmark and security pages say the product is intentionally conservative: it prioritizes low-output guidance, only recommends reducers that proved useful in repeated runs, and treats telemetry tools as measurement-only when that is what the evidence supported.
Who is it meant for?
The site presents Agent Analyzer as a local-first workflow for supported Claude Code, Codex, GitHub Copilot, OpenCode, Cursor, Kiro, Antigravity, and Claude Desktop-related flows, with a plugin built from the sanitized report rather than a generic checklist.
Quick Facts
- Category
- Developer Tool
- Primary workflow
- Local scan of supported agent logs, then sanitized report upload
- Output
- Custom plugin or setup pack based on detected token leaks
- Platforms mentioned
- Claude Code, Codex, GitHub Copilot, OpenCode, Cursor, Kiro, Antigravity, Claude Desktop
- Domain
- analyzer.spec-kitty.ai
- Price
- Not disclosed in the accessible evidence
Alternatives à Agent Analyzer
Ghost
Ghost est un assistant IA en terminal pour discuter, générer du code et lancer des tâches en ligne de commande. Modèles gratuits, Linux, macOS, Windows, open source.
Devin
Devin is an AI coding agent and software engineer that helps developers and engineering teams plan and execute complex software tasks. It is available through desktop, cloud, JetBrains, and CLI surfaces, with plans for individuals, teams, and enterprises.
图像大厨imgcook
imgcook is a design-to-code tool that converts design drafts into front-end code. It supports plugin-based and developer workflows for Sketch, Photoshop, VS Code, and CLI usage.
Pi Coding Agent
Pi Coding Agent is a terminal-based coding agent for developers who want a minimal, extensible harness for interactive work and automation. It supports model switching, session branching, and multiple run modes including TUI, print/JSON, RPC, and SDK.
PromptScout
PromptScout tracks how ChatGPT, Gemini, Google AI Overviews, and Perplexity mention your brand or competitors, then pairs those results with source analysis and website audits. It helps teams decide what to fix in content, positioning, or site readiness next.
Sleek Analytics
Sleek Analytics is a privacy-friendly web analytics tool with real-time visitor tracking, Core Web Vitals, and revenue attribution. It helps site owners understand traffic and conversions without cookie banners or a heavy setup.