Agent Economy ReportAgent Economy ReportSAFE version 1.0.0 · audited 2026-09-17
This package is a text-summarization skill made of instructions rather than executable code. It asks the host agent to create a local folder and store settings, history, saved summaries, and custom templates under ~/.openclaw/summarize-pro/. It does not contain code that sends data over the network or reads credentials or payment details. The only shell command is a one-time mkdir to create the local folder. The publisher's card discloses that local files may retain sensitive pasted content. The observed behavior matches the package's stated purpose, so it can be treated as safe with ordinary care about pasting sensitive material.
| Declared purpose | Text summarization with formats such as TL;DR, bullets, ELI5, executive summaries, action items, meeting notes, email digests, comparisons, multi-language summaries, and custom-length summaries; it claims all processing is local with no external API calls or network requests. |
| Observed behavior | The skill is a prompt-only Markdown file. It instructs the agent to recognize summarization requests, produce summaries in many formats, and store local state under ~/.openclaw/summarize-pro/. It includes one bash command to create that directory, and instructions to read and write settings.json, history.json, saved.json, and templates.json. It also contains attribution and reference links to x.com and clawhub.ai, but no code that performs network requests, accesses credentials, or handles payments. |
| Verdict, rules only | SAFE |
| Verdict, AI | SAFE |
| Final verdict | SAFE (never better than either pass) |
| capability | vs purpose | detail |
|---|---|---|
| shell | needed | The skill tells the agent to run `mkdir -p ~/.openclaw/summarize-pro` for first-run setup. No other shell commands are present. |
| filesystem_read | needed | Reads settings.json, history.json, saved.json, and templates.json from ~/.openclaw/summarize-pro/. |
| filesystem_write | needed | Writes or updates settings.json, history.json, saved.json, and templates.json under ~/.openclaw/summarize-pro/. |
| severity | finding |
|---|---|
| info | Declared local-only behavior The skill description claims processing is local and does not send data to servers. SKILL.md:3 All processing happens locally — NO external API calls, NO network requests, NO data sent to any server. |
| info | Declared tool permissions The skill requests only read and write tools, which matches local file storage needs. SKILL.md:4 metadata: {"clawdbot":{"emoji":"📝","requires":{"tools":["read","write"]}}} |
| low | Shell directory creation The only shell command in the package creates a local directory for the skill's state. SKILL.md:36 mkdir -p ~/.openclaw/summarize-pro |
| info | Local storage path The skill stores all data under a fixed directory in the user's home directory. SKILL.md:66 All data stored under `~/.openclaw/summarize-pro/`: |
| info | Local file boundary claim The skill states it only reads and writes under its own local directory. SKILL.md:76 - Only reads/writes files under `~/.openclaw/summarize-pro/` |
| info | No-network claim The skill instructs that no external API calls or network requests should be made. No code contradicting this was found. SKILL.md:77 - Makes NO external API calls or network requests |
| info | Writes saved summaries The skill can write a summary bookmark to a local JSON file. SKILL.md:459 Save the last summary to `~/.openclaw/summarize-pro/saved.json`: |
| info | Reads summary history The skill reads local summary history for display. SKILL.md:490 Read `history.json` and show recent summaries: |
| low | Automatic local logging Every summary is logged locally in history.json, which could retain sensitive pasted content. SKILL.md:503 Log every summary to history automatically: |
| info | Writes custom templates The skill can save user-defined summary templates locally. SKILL.md:575 Save to `~/.openclaw/summarize-pro/templates.json`. |
| info | External attribution link Contains an attribution link to x.com. This is a Markdown link, not an observed network request. SKILL.md:676 Built by **Manish Pareek** ([@Mkpareek19_](https://x.com/Mkpareek19_)) |
| info | Publisher link Links to the publisher profile on clawhub.ai. This is a reference link, not an observed network request. skill-card.md:9 [mkpareek0315](https://clawhub.ai/user/mkpareek0315) |
| info | Publisher discloses retention risk The skill card discloses that local files may retain sensitive source content or derived summaries. skill-card.md:24 Risk: Local history, stats, saved summaries, and templates may retain sensitive source content or derived summaries under ~/.openclaw/summarize-pro/. |
| info | External skill page link Links to the ClawHub skill page. This is a reference link, not an observed network request. skill-card.md:34 - [ClawHub skill page](https://clawhub.ai/mkpareek0315/skills/summarize-pro) |
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:9 |
| x.com | not in the service index | SKILL.md:676 |
| 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 | 0 |
| Persistence | 0 |
| file | lines | sha256 |
|---|---|---|
SKILL.md | 678 | 59ab8eaf917e5191… |
_meta.json | 6 | e7e78603415fbc98… |
skill-card.md | 52 | be780ff81adbcc49… |
For agents
JSON: https://agenteconomy.report/k/summarize-pro.audit.json · badge: https://agenteconomy.report/k/summarize-pro.audited.svg ·
skill rating: /k/summarize-pro · 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).