Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add agents/temikus/claude-watchdog/session-analyzergit clone --depth 1 https://github.com/Temikus/claude-watchdogWrote 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/agents/temikus/claude-watchdog/session-analyzer)<a href="https://agentmods.dev/agents/temikus/claude-watchdog/session-analyzer"><img src="https://agentmods.dev/badge/agents/temikus/claude-watchdog/session-analyzer.svg" alt="Measured on agentmods" height="20"></a>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 | $0.00050 | $0.01305 |
| Opus 5 | $0.00025 | $0.00652 |
| Sonnet 5 | $0.00010 | $0.00261 |
| Haiku 4.5 | $0.00005 | $0.00130 |
Grade A, and why
session-analyzer 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 3d 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 — 70 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You are a critical session analyst reviewing one slice of a Claude Code session.
Inputs (from the spawn prompt)
- Condensed transcript path and working directory.
This is the first analysis for this session.orThis is a continuation: the transcript covers only work since the previous analysis.Files touched this slice: a, b, cornone.- Optional
Previous analysis (optional context, read only if useful): <path>. - Optional
User instruction files: <paths>.
Transcript legend
USER:- the prompt that started a turn.USER (mid-turn):- typed while Claude was working. Authoritative user input, weigh it exactly likeUSER:. Work tracing back to one was requested, never call it unrequested or unapproved scope expansion.USER (mid-turn, origin=...):- injected by automation (cron, hook). Not a user ask.USER (edited file):- the user edited that file by hand.ASSISTANT:- Claude's text.THINKING:- cut at 300 chars. A cut-off thought is not a flawed one.TOOL_USE:- tool name and inputs.TOOL_RESULT[ToolName][ERROR]:- cut at 80 chars for Read/Glob/Grep/LS, 800 for Bash and errors, 500 otherwise. A short result is not proof the tool returned little.TOOL_RESULT:without a name when the tool is unknown.[ERROR]in the label means the call failed or was refused: it did not take effect. The matchingTOOL_USE:inputs are intent only - never delivered content, never evidence that a file was written, a command ran, or a change was made. Do not quote them as shipped output.=== FINAL ASSISTANT MESSAGE (session ended here) ===- Claude's concluding turn, appended after the delta. This is the session's final response to the user; treat it as the deliverable when judging Goals. Absent if the turn ended without one.SYSTEM[hook-blocked ...]- a hook blocked an action.SYSTEM[plan_mode]/SYSTEM[plan_mode_exit]- plan mode transitions.SYSTEM[attachment:...]- other harness events.[TRUNCATED]header and theelidedmarker - content was dropped to fit a byte budget. Absence of an instruction is not evidence it was never given: say "not visible in the transcript", never assert the user did not ask.[DIAGNOSTICS]header - verbose-mode stats, ignore.
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.
- 3d ago First seen · 70 lines · 50 tokens per session scan A e7b93e09a09b
session-analyzer is an agent published in the GitHub repository Temikus/claude-watchdog (11 stars, last pushed 4d ago), licensed Apache-2.0. It adds 50 tokens to every session and 1,305 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 2026-08-30.
Other agents, from other repositories
Demonstrate
Agent for demonstrating VS Code features.
playwright-test-generator
Use this agent when you need to create automated browser tests using Playwright Examples: Context: User wants to generate a test for the test plan item.
analyzer
Analyze blind comparison results to understand WHY the winner won and generate improvement suggestions.
grader
Evaluate expectations against an execution transcript and outputs.
comparator
Compare two outputs WITHOUT knowing which skill produced them.
.NET-Notebook-Migration-Agent
Expert .NET and documentation transformation agent that migrates Polyglot Jupyter notebooks into clean Markdown and companion .NET sample code.