BB Android
Builds, debugs, ships, and hardens Android apps: Gradle, Jetpack Compose, XML Views, ADB, and the Play release path. Use when a Gradle build fails or a dependency resolves to the wrong version, an APK will not install, or `adb` cannot see the device; when the app crashes, ANRs, leaks, or works on the emulator but not on a real phone; when Compose recomposes too much, state resets on rotation, or a list janks; when WorkManager never runs, an alarm is throttled, or a foreground service crashes at start; when a runtime permission is denied silently, scoped storage blocks a file, or Play rejects a declaration; when raising targetSdk changes behavior; when signing, R8 keep rules, bundle size, staged rollout, or Android vitals are the problem; also Room migrations, deep links, emulators, foldables, and CI builds. Not for Kotlin language mechanics (`kotlin`), store-listing ASO (`google-play-store`), or cross-platform frameworks (`flutter`, `react-native`).
As a process B 70/100 · Nearly there — weak spots: inputs and preconditions, execution cost, running it twice
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 purpose itself is risky: wallets, browser password stores, offensive security. Even an honest implementation gives the agent access to things that cost money.
Explain in the description why the access is needed and how it is limited; add tests that show refusals on dangerous requests.
How to improve
- 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 · 2
✓ No critical or high findings
Medium and low: 2
-
medium Risky intent
intent-wallet-secretsmemory-template.md:65Handles crypto-wallet secrets (seed / mnemonic / private key) — a classic stealer targetIn this domain — **not secrets, keep them**: application id and package name, versionCode and versionName, keystore *alias* and keystore file *path*, certificate SHA-1/SHA-256 fingerprints (they ship
-
medium Risky intent
intent-wallet-secretsrelease.md:25Handles crypto-wallet secrets (seed / mnemonic / private key) — a classic stealer target- The keystore file and its passwords never live in the repository, never in a build file, never in a checked-in properties file. The build reads them from the environment or from a properties file ou
Files scanned: 24. Evidence is masked. Grey chips explain why severity was lowered.
Against the Agent Skills spec
- warning
name-formatname should be kebab-case (lowercase letters, digits, hyphens) - warning
body-longSKILL.md body ≈ 8738 tokens (recommended < 5000); move details to references/ - note
frontmatter-keyunknown frontmatter key "slug" - note
frontmatter-keyunknown frontmatter key "homepage" - note
frontmatter-keyunknown frontmatter key "changelog"
Process rating: all ten parameters 70/100
- 0Inputs and preconditions. Does not say what the process needs to start
- 30Running it twice. 16 mutating operations with no state check
- 40Execution cost. Instruction body is 8738 tokens: crowds the task out of the window
- 60Tools and files. Uses tools (bash) that frontmatter does not declare
- 60Result and completion. Output format stated, no completion criterion
- 100Steps. 42 steps
- 100When it triggers. States when to use and when not to
- 100Failures and branches. 3 branches, has a failure section
- 100Consistency. Name and required fields are in place
- 100Progress reporting. Reports progress
- medium Safety rules and hard prohibitions inside a skill: they belong in the system prompt, here they protect nothing
- low 14 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
- +3Description length 963: 120–800 characters recommended
- +4No input/output examples
- +1No license
- +2Single-language instructions
- +4Description says when NOT to use the skill
- +4Structure: 14 headings
- +3Step-by-step instructions: 42 items
- +3Output format is stated explicitly
Quality base 70; lint remarks subtract, signals add up to 100. Result: 66.