CC deploying-honeytokens-and-canarytokens
Plants Canarytokens-based decoy artifacts (honey credentials, DNS tokens, web-bug URLs, AWS keys, documents, kubeconfigs) using Thinkst's open-source Canarytokens project and alerts via email or webhook when a token is touched. Use for high-fidelity intrusion detection in low-telemetry areas like file shares or credential stores, or to catch credential dumping and data-theft staging.
Plants Canarytokens-based decoy artifacts (honey credentials, DNS tokens, web-bug URLs, AWS keys, documents, kubeconfigs) using Thinkst's open-source…
As a process C 51/100 · Has gaps — weak spots: result and completion, when it triggers, failures and branches
What is at stake
Medium-severity findings: the skill is probably honest, but read what alarmed the scanner.
Below is the worst case for this category. The finding here is medium: the guard saw a sign, not a proof.
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
- 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
✓ No critical or high findings
Medium and low: 6
-
medium Exfiltration
exfil-webhook-urlscripts/agent.py:22Webhook / callback URL commonly used for exfiltration (verify the destination) (quoted — discussed, not commanded)--type aws_keys --webhook https://hooks.slack.com/services/T/B/X \
quoted -
low Risky intent
intent-offensive-securitySKILL.md:3Offensive-security / dual-use content (legitimate for authorised testing; review intended use) (detector / deny-list definition)description: Plants Canarytokens-based decoy artifacts (honey credentials, DNS tokens, web-bug URLs, AWS keys, documents, kubeconfigs) using Thinkst's open-source Canarytokens project and alerts via e
detector -
low Risky intent
intent-offensive-securitySKILL.md:33Offensive-security / dual-use content (legitimate for authorised testing; review intended use) (quoted — discussed, not commanded)This skill maps to MITRE D3FEND's **Decoy File (D3-DF)**, **Decoy User Credential (D3-DUC)**, and **Honeytoken** techniques. From an ATT&CK perspective, a triggered honey credential most commonly evid
quoted -
low Risky intent
intent-offensive-securitySKILL.md:39Offensive-security / dual-use content (legitimate for authorised testing; review intended use)- When seeding decoy credentials into LSASS-reachable memory, browser stores, `.aws/credentials`, or password managers to catch credential dumping and reuse.
-
low Risky intent
intent-offensive-securitySKILL.md:40Offensive-security / dual-use content (legitimate for authorised testing; review intended use)- When instrumenting documents, repos, or wikis to catch data theft and ransomware staging.
-
low Exfiltration
exfil-webhook-urlSKILL.md:123Webhook / callback URL commonly used for exfiltration (verify the destination) (placeholder value)-F 'webhook_url=https://hooks.slack.com/services/T000/B000/XXXX'
placeholder
Files scanned: 5. Evidence is masked. Grey chips explain why severity was lowered.
Against the Agent Skills spec
- note
frontmatter-keyunknown frontmatter key "domain" - note
frontmatter-keyunknown frontmatter key "subdomain" - note
frontmatter-keyunknown frontmatter key "nist_csf" - note
frontmatter-keyunknown frontmatter key "mitre_attack" - note
edit-residuethe text marks something as outdated (lines 155, 157, 158, 160): check that old rules are not kept next to new ones — the full check reads the text for contradictions
Process rating: all ten parameters 51/100
- 0Result and completion. Does not say what the result is
- 0Failures and branches. Linear process with no failure handling
- 0Progress reporting. Says nothing while it works
- 20When it triggers. No condition that starts the skill
- 30Running it twice. 15 mutating operations with no state check
- 60Tools and files. Uses tools (bash, python) that frontmatter does not declare
- 70Inputs and preconditions. Inputs and preconditions are listed
- 100Steps. 24 steps
- 100Consistency. Name and required fields are in place
- 100Execution cost. Instruction body is 3068 tokens
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)
- +3Output format is not stated: the model decides each time
- -4Absolute local paths (C:\Users, /home/…): not portable
- -42 reference files, but SKILL.md never points to them: the model will not open them
- -31 of 1 scripts are never mentioned in SKILL.md
- +2Single-language instructions
- +3Description length 386: enough signal without eating the budget
- +4Structure: 19 headings
- +3Step-by-step instructions: 24 items
- +4Has examples (12 code blocks)
- +1License stated
Quality base 70; lint remarks subtract, signals add up to 100. Result: 71.