Agent Economy ReportAgent Economy ReportCAUTION version 1.0.0 · audited 2026-09-17
This package is a Markdown-only skill that instructs an AI agent how to call an existing Slack tool using JSON action examples. It describes reacting to messages, listing reactions, sending/editing/deleting messages, reading recent messages, managing pins, and fetching member and emoji information. The package does not contain executable code, direct network calls, file writes, obfuscation, or persistence, and it does not collect credentials; it assumes a Slack bot token is already configured. The declared purpose matches the observed behavior. The main caution is that installing this skill could let an agent read and modify Slack content and retrieve member details, so the Slack bot permissions should be reviewed. One metadata file lists shell commands as an output type even though no shell commands are present.
| Declared purpose | Use when you need to control Slack from Clawdbot via the slack tool, including reacting to messages or pinning/unpinning items in Slack channels or DMs. |
| Observed behavior | The package provides metadata and a SKILL.md with JSON examples for Slack tool actions: react, reactions, sendMessage, editMessage, deleteMessage, readMessages, pinMessage, unpinMessage, listPins, memberInfo, and emojiList. No executable code, network calls, file operations, base64, persistence, or shell commands are present; the only external reference is a documentation link to clawhub.ai. |
| Verdict, rules only | SAFE |
| Verdict, AI | CAUTION |
| Final verdict | CAUTION (never better than either pass) |
| capability | vs purpose | detail |
|---|---|---|
| network | needed | The skill is for using the Slack tool, which requires network access to Slack APIs. The package itself contains no direct network code. |
| credentials | needed | SKILL.md line 10 says the Slack tool uses the bot token configured for Clawdbot. The package does not read or store the token itself. |
| severity | finding |
|---|---|
| info | Declared purpose The skill declares that it controls Slack through the existing slack tool. SKILL.md:3 description: Use when you need to control Slack from Clawdbot via the slack tool, including reacting to messages or pinning/unpinning items in Slack channels or |
| low | Relies on preconfigured Slack bot token The skill assumes the Slack bot token is already configured. It does not handle or store credentials. SKILL.md:10 Use `slack` to react, manage pins, send/edit/delete messages, and fetch member info. The tool uses the bot token configured for Clawdbot. |
| medium | Slack message read/send/edit/delete actions enabled The skill instructs that message reading, sending, editing, and deleting are enabled by default. SKILL.md:27 | messages | enabled | Read/send/edit/delete | |
| low | Can send Slack messages Provides an example JSON action for sending messages to a channel or user. SKILL.md:57 "action": "sendMessage", |
| low | Can delete Slack messages Provides an example JSON action for deleting messages by channel and message ID. SKILL.md:77 "action": "deleteMessage", |
| low | Can read recent Slack messages Provides an example JSON action for reading recent messages from a channel. SKILL.md:88 "action": "readMessages", |
| low | Can pin Slack messages Provides an example JSON action for pinning a message in a channel. SKILL.md:98 "action": "pinMessage", |
| low | Can request Slack member info Provides an example JSON action for fetching member information by user ID. SKILL.md:127 "action": "memberInfo", |
| info | Publisher link to clawhub.ai The only external domain is a documentation link to the publisher profile; no hidden network destination was found. skill-card.md:9 [steipete](https://clawhub.ai/user/steipete) |
| low | Metadata lists shell commands as output The metadata mentions shell commands as an output type, but no shell commands appear in the package. This appears to be broader than the actual content. skill-card.md:37 **Output Type(s):** [text, markdown, shell commands, configuration, guidance] |
| info | States no executable code The metadata states the package does not include executable code, which matches the observed Markdown-only content. skill-card.md:43 **Other Properties Related to Output:** [Produces agent-facing instructions for using configured Slack bot permissions; it does not include executable code.] |
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 |
| 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 | 143 | e15de2ab67a9ee49… |
_meta.json | 6 | 0ba4ed14ef7a159c… |
skill-card.md | 51 | bbfd6c002129d29c… |
For agents
JSON: https://agenteconomy.report/k/slack.audit.json · badge: https://agenteconomy.report/k/slack.audited.svg ·
skill rating: /k/slack · 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).