DC meta_ads
Create, write, or manage Meta ads and assets: ad copy, campaigns, spend, reports, audiences, catalogs, product feeds, feed refresh schedules, experiments, and policy. Always load for any request to create or write an ad, or to advertise a product or service, even when no platform is named; this includes sensitive or restricted categories. Always load for any question asking what a Meta, Facebook, or Instagram advertising policy means, allows, prohibits, or requires, including a standalone policy-definition question with no account context. Those questions must use ads_policy_tool, never browser search or memory. Load for a specifically named catalog or feed with an upload or refresh-schedule request; use Ads reads to resolve ownership before writing. Generic unnamed feeds need context. Whether an image, claim or piece of copy may be used in an ad is ALWAYS a Meta Ads task — 'can I use this in an ad', 'is it allowed', 'is this against policy', and any rights, likeness, celebrity, logo or trademark question about advertising with an image, including a follow-up about one just generated. Those are ads-policy questions, not general legal ones. An ad request uses the campaign workflow unless explicitly only an image or organic post.
Create, write, or manage Meta ads and assets: ad copy, campaigns, spend, reports, audiences, catalogs, product feeds, feed refresh schedules, experiments, and…
As a process C 53/100 · Has gaps — References files that are not bundled: <disconnect_url>
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 skill tells the agent to hide things from you: not to show errors, not to mention actions, to report differently from what was done. You lose the ability to see what the agent really did.
Transparency beats a smooth answer. If the goal is to hide technical noise, ask the agent to "summarise briefly", not to "not mention".
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.
- Shorten the description to 1024 characters.
- The SKILL.md body is over 5,000 tokens: move reference detail into references/ and load it when needed.
- The text references files that are not there: add them or drop the references.
- 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 · 1
-
high Concealment
en-hide-from-userreferences/analysis.md:96Instruction to hide actions from the userall at low volume; never tell the user their audience is too broad. If you do
Files scanned: 22. Evidence is masked. Grey chips explain why severity was lowered.
Against the Agent Skills spec
- error
description-longdescription is 1247 chars, limit 1024 - warning
name-formatname should be kebab-case (lowercase letters, digits, hyphens) - warning
body-longSKILL.md body ≈ 6847 tokens (recommended < 5000); move details to references/ - warning
missing-refreference to a missing file: <disconnect_url> - note
frontmatter-keyunknown frontmatter key "title" - note
frontmatter-keyunknown frontmatter key "icon" - note
edit-residuethe text marks something as outdated (lines 211): check that old rules are not kept next to new ones — the full check reads the text for contradictions
Process rating: all ten parameters 53/100
- 0Tools and files. 1 referenced file(s) missing: <disconnect_url>
- 0Result and completion. Does not say what the result is
- 30Running it twice. 12 mutating operations with no state check
- 40Consistency. Frontmatter name (meta_ads) differs from the folder (meta-ads)
- 70Inputs and preconditions. Inputs and preconditions are listed
- 70Execution cost. Instruction body is 6847 tokens
- 85Steps. 28 steps, 1 vague phrases
- 100When it triggers. States when to use and when not to
- 100Failures and branches. 4 branches, has a failure section
- 100Progress reporting. Reports progress
- medium Safety rules and hard prohibitions inside a skill: they belong in the system prompt, here they protect nothing
- low 10 top-level sections: this looks like several domains in one skill
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)
- +3Description length 1247: 120–800 characters recommended
- +3Output format is not stated: the model decides each time
- +1No license
- +2Single-language instructions
- +4Structure: 14 headings
- +3Step-by-step instructions: 28 items
- +4Has examples (5 code blocks)
- +4Reference files are cited in the instructions (18 of 18)
Quality base 70; lint remarks subtract, signals add up to 100. Result: 37.