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 skills add karashiiro/agent-skills --skill analyzing-claude-sessionsgit clone --depth 1 https://github.com/karashiiro/agent-skillsWrote 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/karashiiro/agent-skills/analyzing-claude-sessions)<a href="https://agentmods.dev/skills/karashiiro/agent-skills/analyzing-claude-sessions"><img src="https://agentmods.dev/badge/skills/karashiiro/agent-skills/analyzing-claude-sessions/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/karashiiro/agent-skills/analyzing-claude-sessions"><img src="https://agentmods.dev/badge/skills/karashiiro/agent-skills/analyzing-claude-sessions.svg" alt="Reviewed on agentmods" width="80" 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 5.1 | $0.00041 | $0.01842 |
| Opus 5.5 | $0.00016 | $0.00737 |
| Sonnet 5.5 | $0.00008 | $0.00368 |
| Haiku 4.5 | $0.00004 | $0.00184 |
Grade A, and why
analyzing-claude-sessions 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 2mo 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 — 172 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Analyzing Claude Code Session Logs
Extract insights from Claude Code session transcripts using jq. Session logs are JSONL files where each line is a message in the conversation.
Log Locations
| Log Type | Path |
|---|---|
| Main session | ~/.claude/projects/{project-hash}/{session-id}.jsonl |
| Subagent | ~/.claude/projects/{project-hash}/{session-id}/subagents/agent-{agent-id}.jsonl |
| Tool results | ~/.claude/projects/{project-hash}/{session-id}/tool-results/ |
Find the most recent session:
ls -t ~/.claude/projects/*/*.jsonl | head -1
List subagents for a session:
ls ~/.claude/projects/{project-hash}/{session-id}/subagents/
Message Structure
Each JSONL line has this shape:
{
"type": "user|assistant|progress|system|file-history-snapshot",
"message": {
"role": "user|assistant",
"content": "string" | [{"type": "text|tool_use|tool_result", ...}],
"model": "claude-sonnet-4-5-...",
"usage": {"input_tokens": N, "output_tokens": N, ...},
"stop_reason": "end_turn|tool_use|..."
},
"timestamp": "ISO8601",
"uuid": "...",
"agentId": "...",
"sessionId": "..."
}
Assistant messages have content as an array of blocks:
{"type": "text", "text": "..."}— text output{"type": "tool_use", "name": "Read", "id": "toolu_...", "input": {...}}— tool call{"type": "tool_result", "tool_use_id": "toolu_...", "content": "...", "is_error": bool}— tool response
Shell Quoting Note
jq filters work fine inline as long as the file path is either a literal or expanded in the same && chain as the variable assignment. The select(.type == "tool_use") syntax with inner double quotes works correctly inside single-quoted jq filters.
For array content blocks, use the if type == "array" guard since user messages have string content:
jq -r '.message.content | if type == "array" then [.[] | select(.type == "tool_use") | .name] | .[] else empty end' session.jsonl
For loops over subagent files, assign the directory path first:
SUBS="path/to/subagents" && for f in "$SUBS"/agent-*.jsonl; do
jq -r '...' "$f"
done
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.
- 2mo ago First seen · 172 lines · 41 tokens per session scan A 10fb1ccf36ca
analyzing-claude-sessions is a skill published in the GitHub repository karashiiro/agent-skills (2 stars, last pushed 6mo ago), licensed Unlicense. It adds 41 tokens to every session and 1,842 once invoked, about $0.0002 per session on Opus 5.5. 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.
analyzing-claude-sessions compared with skills from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| analyzing-claude-sessionskarashiiro/agent-skills | Skill | 2 | 6mo ago | Unlicense | A | 41 | Claude Code, Codex |
| imagegenopenai/codex | Skill | 134k | today | Apache-2.0 | A | 113 | Codex |
| local-ai-agentsmicrosoft/ai-agents-for-beginners | Skill | 79k | today | MIT | A | 200 | Claude Code, Codex |
| playwright-climicrosoft/playwright | Skill | 99k | today | Apache-2.0 | A | 19 | Claude Code |
| systematic-debuggingobra/superpowers | Skill | 314k | 3d ago | MIT | A | 21 | Claude Code |
| next-cache-components-adoptionvercel/next.js | Skill | 143k | today | MIT | A | 95 | Claude Code |
| chat-pet-sprite-creationmicrosoft/vscode | Skill | 195k | today | MIT | A | 53 | Claude Code, Codex |