Agent Economy ReportAgent Economy ReportSAFE version 1.0.0 · audited 2026-09-09
This is a documentation-only skill: it contains no executable code, just a Markdown guide telling the agent where its own conversation logs are stored and giving example shell one-liners (using jq, rg, head, awk, sort) to search them. The commands only read files under the local `~/.clawdbot/agents/<agentId>/sessions/` directory and print counts, dates, costs, tool usage, and matching text. There is no network activity, no credential access, no installation, no persistence, and no obfuscated or encoded content; the only URLs in the package are the publisher's ClawHub profile and skill page in the descriptive card. The main thing to be aware of is that the skill's whole purpose is to open full transcripts of past conversations, which can contain private information or secrets that were pasted into chats, and it also pulls arbitrary past text back into the agent's context. Behavior matches the stated purpose with nothing exceeding it.
| Declared purpose | Search and analyze the agent's own local session logs (older/parent conversations) using jq and ripgrep, to recover prior context and summarize message, cost, and tool-usage statistics. |
| Observed behavior | Pure Markdown instructions. Points the agent at ~/.clawdbot/agents/<agentId>/sessions/ and supplies read-only bash snippets (for-loops with head, jq, ls, awk, sort, uniq, rg) to list sessions, filter by date, extract user/assistant text, sum costs, count messages/tokens, break down tool calls, and grep across all session files. No writes, no network calls, no environment variable or credential reads, no downloads or execution of remote code. |
| Verdict, rules only | SAFE |
| Verdict, AI | SAFE |
| Final verdict | SAFE (never better than either pass) |
| capability | vs purpose | detail |
|---|---|---|
| filesystem_read | needed | Reads local session transcripts and the sessions.json index under `~/.clawdbot/agents/<agentId>/sessions/` (SKILL.md lines 17, 35, 44, 66, 91, 104). These files hold complete past conversations, so this is broad access to potentially sensitive content. |
| shell | needed | All examples are bash snippets using jq, rg, head, ls, awk, sort, uniq (SKILL.md line 34 onward). The commands shown are read-and-print only; none modify or delete files. |
| other | needed | Declares required binaries jq and rg in metadata (SKILL.md line 4); it does not attempt to install them. |
| severity | finding |
|---|---|
| low | Reads full local conversation history The skill directs the agent to the directory containing complete session transcripts. This is required for the declared purpose, but those transcripts may include private data or secrets previously pasted into chats. All access is local; nothing is transmitted anywhere in this package. SKILL.md:17 Session logs live at: `~/.clawdbot/agents/<agentId>/sessions/` (use the `agent=<id>` value from the system prompt Runtime line). |
| info | Cross-session full-text search Suggests grepping every session file at once. Read-only, but it broadens exposure from one conversation to all stored conversations for that agent. SKILL.md:91 rg -l "phrase" ~/.clawdbot/agents/<agentId>/sessions/*.jsonl |
| info | External binaries required, not installed Declares a dependency on jq and rg. The skill does not download or install anything; if the tools are missing the commands simply fail. SKILL.md:4 metadata: {"openclaw":{"emoji":"📜","requires":{"bins":["jq","rg"]}}} |
| info | Only network reference is the registry page The single external domain in the package is clawhub.ai, appearing as a publisher/reference link in the descriptive card. No code fetches it. skill-card.md:7 [guogang1024](https://clawhub.ai/user/guogang1024) <br> |
0 AI finding(s) were dropped because their file, line or quote did not match the package.
| domain | service rating | where |
|---|---|---|
| clawhub.ai | not in the service index | skill-card.md:7 |
| Binaries invoked | bash |
| Environment variables read | none |
| Hard-coded wallets | none |
| Pipes a download to a shell | 0 |
| eval / exec / subprocess | 0 |
| base64 blobs | 0 |
| File writes | 6 |
| Persistence | 0 |
| file | lines | sha256 |
|---|---|---|
SKILL.md | 105 | f87582f99fee9009… |
_meta.json | 6 | ebc38e96c988b61e… |
skill-card.md | 37 | f14238524bef749d… |
For agents
JSON: https://agenteconomy.report/k/session-logs.audit.json · badge: https://agenteconomy.report/k/session-logs.audited.svg ·
skill rating: /k/session-logs · commission an audit of any skill: US$ 29 per version.
The complete published package of this exact version was downloaded from the registry and read statically; nothing was executed. A deterministic pass extracts network destinations, binaries, environment variables, writes, install commands, obfuscation markers, subprocesses and wallets, each with file and line. An AI then reads the whole package with those facts and writes the summary, the capabilities and the findings under a strict schema; every finding it produces must cite a file, a line and the exact text, or it is dropped. The final verdict is the worst of the two passes. The audit does not change the skill's trust tier (policy); the author may respond through the dispute channel and the response is published here. Commissioned by: the Agent Economy Report (free program: the 150 most downloaded skills, September 2026).