agentleFS
apexbymanishapexbymanish/claude-ai-agents-ios/.claude/skills/ios-evidence-reporting/SKILL.md

ios-evidence-reporting

Skill0 starsChanged 3 months agoMITView on GitHub

Standard for reporting task outcomes with verifiable evidence instead of confident prose. Use before writing "it works", "done", "fixed", "faster", "thread-safe", "secure", or "production-ready" at the end of any iOS agent task — implementation, testing, review, audit, or architecture consultation.

The file

We split this skill into the rules it gives an agent

6 rules. About 0 are common, 5 are distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈5

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    If an agent can't reach the tier a claim needs, it reports the tier it actually reached and marks the gap — it never rounds up.

    line 114
  2. Distinctive:

    Do not write these words in a final report unless the matrix above is satisfied for that exact claim:

    line 119
  3. Distinctive:

    Use precise language instead:

    line 139
  4. Distinctive:

    Not a replacement for the full prose report each agent already produces — it's a compact summary at the end, and every line in it must be …

    line 205
  5. Very specific:

    Not a demand to run tools you don't have.

    line 208
  6. Distinctive:

    Not a reason to invent categories.

    line 211

≈ “Common” is an estimate: at least 1 in 5 of the 372,429 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

iOS Evidence-Based Reporting

AI confidence is not evidence. AI reasoning is not runtime evidence. A code change is not automatically a verified fix. Every claim in a report is backed by evidence at a specific tier — and a claim never gets reported at a higher tier than the evidence actually reached.

The seven evidence tiers

From weakest to strongest. A claim's tier is the weakest link that actually backs it — gathering strong evidence for half a claim doesn't license reporting the whole claim at that strength.

  1. ASSUMPTION — a belief based on reasoning, not evidence. Never reported as verified. Phrase it as a hypothesis: "likely a decoding cost" not "is a decoding cost."
  2. STATIC_ANALYSIS — evidence from reading source, a compiler diagnostic, AST/dependency inspection, or a static-analysis tool (including an MCP analyzer such as ios-agent-mcp) run over source without building or running the app. Reading the output of grep, wc, or git diff also lands here — running a text command over source is still static analysis, not runtime evidence, regardless of whether the agent doing it is otherwise read-only. Example: "no obvious retain cycle found during static inspection" — this does not mean "no memory leak exists."
  3. BUILD_VERIFIED — the project actually compiled. Example: "xcodebuild completed successfully."
  4. TEST_VERIFIED — an automated test suite ran and its result is shown. Example: "42 tests passed."
  5. RUNTIME_VERIFIED — the app was actually launched, installed, or interacted with (simulator or device) and its behavior observed — a screenshot, a deep link reaching the right screen, a manual walkthrough. Confirms behavior, not a number — see the next tier for that.
  6. RUNTIME_MEASURED — an actual number came from actually running the app: Instruments (Allocations, Leaks, Time Profiler), MetricKit, XCTMetric, os_signpost, a launch-time or memory measurement. Required for any performance or memory improvement claim — no lower tier can support one.
  7. HUMAN_VERIFICATION — the agent cannot perform the verification this claim needs (driving the Instruments GUI, for instance) and says so explicitly, rather than guessing or silently skipping it.

Tiers 3-6 all require something to have actually run; only 5 and 6 require the app itself to have run. Static analysis — including an MCP tool's structured findings — never outranks build/test/runtime tiers, no matter how sophisticated the analysis.

MCP/tool-assisted analysis is still whatever tier it actually is

Canonical rule, referenced rather than re-derived: an MCP server's output is graded by what it actually did, not by the fact that a tool (rather than a human) did it.

  • ios-agent-style servers are static analyzers — they read source and never build or run the app. Their findings are STATIC_ANALYSIS, identical tier to a manual code read, never higher, no matter how structured or confident-looking their output is.
  • ios-simulator-style servers actually build, install, and launch the app — their build_project/run_tests/launch_app/screenshot output can genuinely earn BUILD_VERIFIED, TEST_VERIFIED, or RUNTIME_VERIFIED, because the app actually ran. It still cannot produce RUNTIME_MEASURED on its own unless it reports an actual measured number (a screenshot confirms behavior, not a quantity).

A pattern match is a lead, not a finding

Canonical rule, referenced rather than re-derived: a grep hit, an MCP tool flag, or any other pattern match is a candidate to verify, never a confirmed issue on its own — regardless of which agent produced it or how narrow the pattern looks. "UserDefaults called near the word token" is a lead; "this stores the session token insecurely" requires having actually traced that it's a real token, on that path, with no Keychain wrapper already handling it. Report a match as a lead — cite the evidence, state what would confirm it — until the trace is actually done. This applies most often in ios-security-review, ios-legacy-mapping, and ios-feature-implementation's Security Review, but the rule itself lives here rather than being restated in each.

Independent review requirement

Any report reaching BUILD_VERIFIED or higher — not just STATIC_ANALYSIS — is handed to ios-evidence-reviewer before being presented as final. The agent that ran the build/test/measurement is not the sole authority on whether its own report is honest about it; see ios-evidence-reviewer for what it checks. Skip this only for a report making no claim above STATIC_ANALYSIS/ASSUMPTION (a pure read-only consult has nothing for the reviewer to add).

Claim → minimum evidence matrix

ClaimMinimum evidence
Code compilesBUILD_VERIFIED
Unit/UI tests passTEST_VERIFIED
A feature's logic/behavior is correctTEST_VERIFIED (the tests that exercise it)
The UI renders/looks/behaves correctlyRUNTIME_VERIFIED — a passing unit test cannot confirm a visual or layout claim; something has to actually be seen running
An API is available at the stated minimumSTATIC_ANALYSIS (availability check)
No obvious retain cycleSTATIC_ANALYSIS only — never "no leak exists"
No runtime leak observedRUNTIME_MEASURED (Instruments/Memory Graph)
Memory usage improvedRUNTIME_MEASURED, before/after
Performance improvedRUNTIME_MEASURED, before/after
Thread-safeSTATIC_ANALYSIS + concurrency-relevant tests — never inspection alone
Architecture preserved / consistentSTATIC_ANALYSIS (architecture review)
Security issue addressedSTATIC_ANALYSIS/security review, plus a regression test where practical
App Store readySTATIC_ANALYSIS (code-visible checks) — metadata/screenshots still need HUMAN_VERIFICATION

If an agent can't reach the tier a claim needs, it reports the tier it actually reached and marks the gap — it never rounds up.

Never say, unless the evidence supports it

Do not write these words in a final report unless the matrix above is satisfied for that exact claim:

  • "fixed" / "working" — unless RUNTIME_VERIFIED or TEST_VERIFIED backs it.
  • "optimized" / "faster" / "uses less memory" — unless RUNTIME_MEASURED before/after backs it.
  • "leak-free" — never claimable from STATIC_ANALYSIS alone.
  • "thread-safe" — never claimable from STATIC_ANALYSIS alone.
  • "secure" — a security review can say "no issues found in this review's scope," never an unqualified "secure."
  • "iOS [version] compatible" — unless an availability check (STATIC_ANALYSIS at minimum) actually ran.
  • "production-ready" — never claimable by any single agent. It's a cross-cutting claim spanning build, tests, security, performance, accessibility, and App Store readiness at once; no one specialist's evidence covers all of it. State which specific checks passed instead ("build verified, tests pass, no security findings in this review's scope") and leave the umbrella claim unmade.

Use precise language instead:

Instead ofSay
"Fixed it.""Implemented the proposed fix."
"Build works.""Build verified successfully."
"No memory leak.""Static inspection found no obvious retain cycle."
"It's faster now.""Memory/performance improvement was not measured." (if it wasn't)
"Should be fine at runtime.""Runtime verification is still required."
"This is production-ready.""Build/tests/security review passed; App Store readiness and UX review are separate checks not yet run."

The status block

End every report with a compact status block: one line per category relevant to your domain, each marked:

  • ✓ <TIER> — checked and fine, backed by evidence at <TIER>.
  • ✗ <TIER> — checked, and a real problem was found, backed by evidence at <TIER>.
  • ⚠ — not checked, or only an ASSUMPTION exists — say why, and what tier would actually verify it. Never silently drop a category that matters because it wasn't checked; mark it ⚠, don't omit it.

Pick categories from your own agent's checklist. Don't force categories that don't apply to your role — a read-only review agent has no BUILD line; an architecture consult has no TEST line.

Example: an implementation/verification task

(ios-feature-implementation, ios-unit-test-engineer, ios-ui-test-engineer, ios-memory-performance-engineer)

BUILD        ✓ BUILD_VERIFIED     xcodebuild succeeded
TEST         ✓ TEST_VERIFIED      24/24 tests passed
AVAILABILITY ✓ STATIC_ANALYSIS    iOS 14 compatible (checked 2 new API calls)
MEMORY       ✓ STATIC_ANALYSIS    no retain-cycle pattern found
             ⚠ HUMAN_VERIFICATION Instruments not run — recommend an Allocations pass
PERFORMANCE  ✓ STATIC_ANALYSIS    no blocking main-thread work introduced
SECURITY     ✓ STATIC_ANALYSIS    token stored via existing Keychain wrapper
DIFF         ✓ STATIC_ANALYSIS    4 files changed, diff reviewed end to end

Example: a read-only review/audit/consult task

(ios-architect, ios-ux-reviewer, ios-legacy-auditor)

ACCESSIBILITY   ✗ STATIC_ANALYSIS  1 tap target below 44pt (ProfileView.swift:82)
CONVENTIONS     ✓ STATIC_ANALYSIS  matches platform navigation patterns
CONSISTENCY     ✓ STATIC_ANALYSIS  reuses existing button style throughout
ARCHITECTURE    ✓ STATIC_ANALYSIS  MVC-in-name-only (38 VCs >400 lines via grep,
                210 .shared refs via grep)
TESTABILITY     ⚠                  proposed design not yet unit-tested — no test exists
SECURITY        ⚠                  3 leads to verify — see Security signals section

Note ARCHITECTURE above is STATIC_ANALYSIS, the same tier as CONVENTIONS — running grep/wc and reading real command output is still reading source, not running the app, even for a read-only agent whose whole job is that kind of read. Don't grade a grep-derived count any higher than a plain code read; the strength that matters is static vs. runtime, not manual vs. tool-assisted.

What this is not

  • Not a replacement for the full prose report each agent already produces — it's a compact summary at the end, and every line in it must be traceable to something stated in the body above it.
  • Not a demand to run tools you don't have. A read-only agent marks ⚠ (or HUMAN_VERIFICATION for a specific claim) honestly rather than fabricating a check it couldn't perform.
  • Not a reason to invent categories. If nothing in your domain checklist is relevant to a category, leave it out — don't pad the block to look thorough.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.