Agent Economy ReportAgent Economy ReportCAUTION version 1.0.0 · audited 2026-09-09
This package is a text-only "marketing mode" persona and knowledge base. It contains prompt/instruction Markdown files (marketing frameworks, SEO checklists, psychology models, copywriting patterns) plus metadata. There is no script, no code execution, no file writing, no network calls, and no access to credentials or payments anywhere in the files. The only capability worth reviewing is the metadata declaring an optional global npm install of an external package (`npm install -g @thesethrose/marketing-mode`), which would run third-party install code that is not included here and therefore cannot be audited. Nothing in the text tries to override the agent's other rules or redirect data or money.
| Declared purpose | A combined marketing skill/persona ("Mark the Marketer") providing marketing strategy, copywriting, SEO, conversion optimization, and paid-advertising guidance. |
| Observed behavior | All files are documentation and prompt text that shape how the agent talks about marketing. The agent is told to ask clarifying questions and recommend tactics. Metadata declares node/npm requirements and an install entry pointing at a global npm package; no command is executed by the package itself. |
| Verdict, rules only | SAFE |
| Verdict, AI | CAUTION |
| Final verdict | CAUTION (never better than either pass) |
| capability | vs purpose | detail |
|---|---|---|
| install_packages | exceeds purpose | skill.json line 25 declares `"npm": "npm install -g @thesethrose/marketing-mode"` and lines 42-49 define an npm install step. A purely text-based advice skill needs no package install; installing a global npm package runs unaudited third-party install scripts. |
| network | exceeds purpose | No network calls in any file. The only URLs are documentation/reference links (clawhub.ai in skill-card.md line 7, github.com repository in skill.json line 8). |
| severity | finding |
|---|---|
| medium | Declares a global npm install of an external, unaudited package The skill's metadata offers an install path that installs a package globally. npm install can execute lifecycle scripts from the remote package, which is code not present in this package and therefore not reviewable here. A text-only marketing advice skill does not require this. Prefer using only the bundled Markdown files. skill.json:25 "npm": "npm install -g @thesethrose/marketing-mode" |
| low | Install step points to remote package The clawdbot install block defines an 'npm' kind install step for the same remote package, so an agent honoring this metadata could fetch and install remote code without further review. skill.json:46 "package": "@thesethrose/marketing-mode", |
| info | Requires npm/node although no code is shipped The skill declares node and npm as requirements (SKILL.md lines 15-16) but ships no executable files, so the requirement is unnecessary for the documented behavior. SKILL.md:16 npm: true |
| info | Publisher discloses the npm install risk The skill card itself acknowledges that the global npm install path can execute package code, and advises confirming trust in the external package. This matches the audit finding and is a point in the publisher's favor for transparency. skill-card.md:20 Risk: One install path uses a global npm package, which can run package code during installation. <br> |
| info | Persona instructions only The prompt file changes the agent's tone and framing for marketing questions. It requests no tools, credentials, network access, or file operations. mode-prompt.md:3 You are Mark the Marketer, a growth-obsessed marketing strategist. |
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 |
| github.com | not in the service index | skill.json:8 |
| Binaries invoked | bash, node, npm |
| 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 |
|---|---|---|
README.md | 49 | b4397c9a88f74294… |
SKILL.md | 693 | 27f1838f183c9593… |
_meta.json | 6 | 26784ae7e4e8dd53… |
mode-prompt.md | 39 | e6260530dfc9d274… |
skill-card.md | 40 | 8bf224f78128b5c9… |
skill.json | 51 | 148a832812f3677a… |
For agents
JSON: https://agenteconomy.report/k/marketing-mode.audit.json · badge: https://agenteconomy.report/k/marketing-mode.audited.svg ·
skill rating: /k/marketing-mode · 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).