Agent Economy ReportAgent Economy ReportCAUTION version 1.0.0 · audited 2026-09-17
Blogwatcher is a Markdown skill card that tells an agent to install the blogwatcher Go CLI from GitHub and use it to add blogs, scan RSS/Atom feeds, list articles, and mark articles read. The package files contain instructions and metadata only; no executable payload, obfuscated code, credential collection, payment redirection, or persistence commands appear in the provided files. The declared operation requires shell execution of the blogwatcher command and network access to feed URLs. The main concern is that the install instructions use an unpinned @latest Go module, so the fetched executable could change after review. The package itself acknowledges this risk and recommends pinning, but its own install metadata still uses @latest.
| Declared purpose | Monitor blogs and RSS/Atom feeds for updates using the blogwatcher CLI. |
| Observed behavior | The package is a short instruction card. It directs the agent to install the blogwatcher Go binary from github.com/Hyaxia/blogwatcher using go install ...@latest, then run commands such as add, blogs, scan, articles, read, read-all, and remove. It includes example feed URLs example.com and xkcd.com. No direct file writes, environment variable access, persistence, or payment wallet addresses are present in the package files. |
| Verdict, rules only | SAFE |
| Verdict, AI | CAUTION |
| Final verdict | CAUTION (never better than either pass) |
| capability | vs purpose | detail |
|---|---|---|
| install_packages | needed | SKILL.md line 5 metadata provides a Go install module github.com/Hyaxia/blogwatcher/cmd/blogwatcher@latest, and line 13 repeats the go install command. |
| shell | needed | The skill instructs running the blogwatcher CLI commands, for example SKILL.md line 16 `blogwatcher --help` and line 21 `blogwatcher scan`. |
| network | needed | The installed CLI needs to fetch RSS/Atom feeds and the Go module. Examples include https://example.com at SKILL.md line 19 and https://xkcd.com at line 33; the homepage links to github.com at line 4. |
| severity | finding |
|---|---|
| medium | Unpinned mutable installation source The install metadata uses @latest for the Go module, so `go install` can fetch different code over time. The package therefore installs a third-party executable without a reviewed version pin. SKILL.md:5 "module":"github.com/Hyaxia/blogwatcher/cmd/blogwatcher@latest" |
| low | Manual install command also uses @latest Repeats the same unpinned install target as a shell command. It matches the declared purpose but should be pinned before use in a sensitive environment. SKILL.md:13 - Go: `go install github.com/Hyaxia/blogwatcher/cmd/blogwatcher@latest` |
| low | Add command accepts remote URL The skill instructs the CLI to track a user-supplied URL, which means the installed binary will make network requests to that URL. Feed URLs should be restricted to trusted sources. SKILL.md:19 blogwatcher add "My Blog" https://example.com |
| info | Shell command for feed scanning This is a documented CLI command for the agent to run. Shell/process execution is necessary for the declared blog-monitoring purpose. SKILL.md:21 - Scan for updates: `blogwatcher scan` |
| info | Self-disclosed mutable module risk The card acknowledges the same supply-chain risk as the @latest install target in SKILL.md line 5. skill-card.md:24 Risk: Installing the CLI from a mutable Go module version can fetch code that has not been reviewed for this release. |
| low | Self-disclosed risk of inherited secrets The card warns that the installed CLI could inherit agent environment secrets or filesystem access. No actual secret collection was observed in the package, but this reinforces running the CLI with minimal privileges. skill-card.md:28 Risk: The installed CLI may inherit unnecessary secrets or sensitive filesystem access from the agent environment. |
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 |
| example.com | not in the service index | SKILL.md:19 |
| github.com | not in the service index | SKILL.md:4 |
| xkcd.com | not in the service index | SKILL.md:33 |
| Binaries invoked | none |
| 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 | 46 | c22cd0c2c61891df… |
_meta.json | 6 | 9a7fbf7b383b4d72… |
skill-card.md | 53 | 66301d6bf76e1ffd… |
For agents
JSON: https://agenteconomy.report/k/blogwatcher.audit.json · badge: https://agenteconomy.report/k/blogwatcher.audited.svg ·
skill rating: /k/blogwatcher · 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).