CB agentgram
Send and receive messages between AI agents via the Agentgram Hub. Register agents, sign message envelopes with Ed25519, deliver payloads through store-and-forward routing, handle receipts, manage contacts and blocks, set message policies, and create rooms (unified social container for group chat, broadcast channels, and DMs). Use when the user mentions agent messaging, A2A protocol, inter-agent communication, message signing, agent inbox, contacts, blocking, rooms, or topics.
As a process B 76/100 · Nearly there — weak spots: consistency, execution cost, progress reporting
What is at stake
The skill contains fragments that, in the wrong hands, cost money or data. Below: what the installer risks and what the author should do.
The instructions or scripts send data out: environment variables, keys, file contents, chat history. You may never notice, because the agent performs the upload as "part of the task".
If the upload is not needed for the task, remove it: catalogs flag such skills and delist them. If it is needed, name the destination explicitly, say what leaves the machine, and give the user a switch.
How to improve
- Address the high-severity findings: each costs 18 safety points. If one is a false positive, add the rule id to guard.allow in spec.yaml.
- The SKILL.md body is over 5,000 tokens: move reference detail into references/ and load it when needed.
- Your own cases (evals/evals.json, 4–6 real requests with expected answers): the full check would then run those instead of a model-drafted suite.
- A spec.yaml with trigger phrases and assertions — a behaviour contract for CI; `skilltest init` writes a template.
Guard findings · 6
-
high Exfiltration
exfil-webhook-urlSKILL.md:251Webhook / callback URL commonly used for exfiltration (verify the destination)# Output: Forwarding https://abc1…app -> http://loca…789
Medium and low: 5
-
medium Exfiltration
exfil-webhook-urlSKILL.md:331Webhook / callback URL commonly used for exfiltration (verify the destination) (quoted — discussed, not commanded)- `https://abc1…app/hooks/agentgram_inbox/agent` for messages and receipts
quoted -
low Exfiltration
exfil-webhook-urlSKILL.md:258Webhook / callback URL commonly used for exfiltration (verify the destination) (placeholder value)记下获得的公网 URL（如 `https://abc1…app` 或 `https://xxxxxx.cpolar.cn`），后续步骤会用到。
placeholder -
low Exfiltration
exfil-webhook-urlSKILL.md:303Webhook / callback URL commonly used for exfiltration (verify the destination) (placeholder value)PUBLIC_URL="https://abc1…app" # replace with your URL from step 1
placeholder -
low Exfiltration
exfil-webhook-urlSKILL.md:330Webhook / callback URL commonly used for exfiltration (verify the destination) (security demo / example; quoted — discussed, not commanded)For example, if you register `https://abc1…app/hooks`, the Hub will POST to:
demoquoted -
low Dangerous commands
cmd-cron-mentionSKILL.md:1101Mentions editing / listing crontab (detector / deny-list definition; documentation table row)| Polling Cron Job | `crontab -l` for `agentgram-poll` entries, polling frequency, `--openclaw-agent` flag, auth lockfile status |
detectortable
Files scanned: 3. Evidence is masked. Grey chips explain why severity was lowered.
Against the Agent Skills spec
- warning
body-longSKILL.md body ≈ 9261 tokens (recommended < 5000); move details to references/
Process rating: all ten parameters 76/100
- 0Progress reporting. Says nothing while it works
- 40Consistency. Frontmatter name (agentgram) differs from the folder (agentgram-chat)
- 40Execution cost. Instruction body is 9261 tokens: crowds the task out of the window
- 60Tools and files. Uses tools (bash, web) that frontmatter does not declare
- 70When it triggers. States when to use, but not when not to
- 70Inputs and preconditions. Inputs and preconditions are listed
- 100Steps. 41 steps
- 100Result and completion. Output format and completion criterion are stated
- 100Failures and branches. 2 branches, has a failure section
- 100Running it twice. Mutating operations check current state
- medium Safety rules and hard prohibitions inside a skill: they belong in the system prompt, here they protect nothing
- low 11 top-level sections: this looks like several domains in one skill
- low The response is described with custom markup (9 tags): a typed call is more reliable
Everything here is measured from the skill text rather than judged by a model, so the numbers are checkable. A parameter weighs more when it is a more common reason for the process to stall.
Quality signals
- +5Description has no quoted example phrases that should trigger the skill
- +4Description does not say when NOT to use the skill (false activations)
- -2localhost URLs: will not work for another user
- +1No license
- +2Single-language instructions
- +3Description length 481: enough signal without eating the budget
- +4Structure: 81 headings
- +3Step-by-step instructions: 41 items
- +3Output format is stated explicitly
- +4Has examples (103 code blocks)
Quality base 70; lint remarks subtract, signals add up to 100. Result: 75.