Agent Economy ReportAgent Economy ReportSAFE version 1.1.3 · audited 2026-09-09
This is a text-only skill: a set of Markdown instructions and nine weekly-report templates (engineer, manager, team, sales, marketing, project, HR, operations, finance). It tells an AI assistant to ask the user for their job role and weekly work details, then fill in the matching template and offer edits. There is no executable code, no network calls, no file access to your system, and no credential or payment handling anywhere in the package. The only shell commands mentioned (chmod, bash, git) appear in the README's instructions for the skill's own maintainers on how to publish new versions, not in anything the agent runs. The README also warns users not to paste confidential or personal data into an AI service. Nothing in the package attempts to override the agent's other rules or send data anywhere.
| Declared purpose | A prompt-only skill that converts natural-language descriptions of a person's week into a structured, role-specific weekly report, with follow-up tone/detail adjustments. |
| Observed behavior | Behavior matches the declared purpose exactly: role identification, information collection questions, template selection, report generation, and revision prompts. All content is static Markdown; no scripts, no data collection, no external endpoints besides the publisher's own ClawHub reference links. |
| Verdict, rules only | SAFE |
| Verdict, AI | SAFE |
| Final verdict | SAFE (never better than either pass) |
| capability | vs purpose | detail |
|---|---|---|
| other | needed | Pure prompt/instruction content only. The skill defines conversational behavior (ask role, ask for tasks/blockers/plans, emit a Markdown table) and requests no system capabilities. |
| shell | exceeds purpose | README.md lines 142-148 document maintainer-only publishing commands (chmod +x scripts/clawhub-publish.sh, ./scripts/clawhub-publish.sh, clawhub publish). The referenced script is not included in the package and nothing instructs the agent to execute these; they are human documentation. |
| network | exceeds purpose | The only URLs are informational links to clawhub.ai publisher and skill pages (skill-card.md lines 7 and 27). No fetch, curl, or upload behavior is defined. |
| severity | finding |
|---|---|
| info | Maintainer publishing commands documented in README The README documents shell commands for the skill's own release process (chmod, running clawhub-publish.sh, git cleanliness check). The script itself is not shipped in the package (line 155 notes scripts/ is ignored at publish time), and nothing directs the agent to run these. Informational only, but a user should not blindly execute commands from a skill's README. README.md:144 chmod +x scripts/clawhub-publish.sh |
| info | No data collection claim is consistent with the files The package contains no code, no network requests and no filesystem operations, so the claim that it neither collects nor transmits data is consistent with what is present. The only writes recorded by static analysis (SKILL.md lines 29 and 45) are quoted question text the assistant should say to the user, not disk writes. README.md:126 本 Skill 只是一个 Prompt 模板，**本身不会收集、存储或传输任何数据** |
| info | Only external domain is the registry itself The single external domain referenced anywhere in the package is clawhub.ai, used for publisher and skill listing links in the skill card. There are no hidden or unexpected network destinations. skill-card.md:7 [wscats](https://clawhub.ai/user/wscats) <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, chmod, git |
| Environment variables read | none |
| Hard-coded wallets | none |
| Pipes a download to a shell | 0 |
| eval / exec / subprocess | 0 |
| base64 blobs | 0 |
| File writes | 2 |
| Persistence | 0 |
| file | lines | sha256 |
|---|---|---|
CHANGELOG.md | 27 | 9d31427f317d11ef… |
README.md | 176 | 5f0365325ee90e99… |
SKILL.md | 420 | 7e89b894e7f8eb03… |
_meta.json | 6 | ee0898f6a068b8db… |
skill-card.md | 43 | dbb47952ae63b5dc… |
For agents
JSON: https://agenteconomy.report/k/smart-weekly-report.audit.json · badge: https://agenteconomy.report/k/smart-weekly-report.audited.svg ·
skill rating: /k/smart-weekly-report · 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).