agentleFS
apexbymanishapexbymanish/claude-ai-agents-ios/.claude/skills/ios-legacy-mapping/SKILL.md

ios-legacy-mapping

Skill0 starsChanged 3 months agoMITView on GitHub

Step-by-step procedure for mapping an undocumented, unfamiliar, or large legacy iOS codebase (Objective-C, UIKit, SwiftUI, or a mix) before making changes. Use when onboarding onto a project with no CLAUDE.md, no architecture docs, or when asked to "explore this codebase" or "figure out how this app is structured".

The file

We split this skill into the rules it gives an agent

1 rule. About 0 are common, 1 is distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈1

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Write findings as a CLAUDE.md-ready summary with these sections:

    line 122

≈ “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 Legacy Codebase Mapping

A read-only reconnaissance procedure. Do not modify code while running this — the output is a document, not a refactor.

1. Inventory targets and schemes

find . -name "*.xcodeproj" -o -name "*.xcworkspace"
xcodebuild -list -project <Project>.xcodeproj   # or -workspace

Note every target (app, extensions, frameworks, test targets) and which scheme builds what. A large app often has more targets than the obvious main one (share extensions, widgets, watch app).

2. Detect the architecture actually in use — don't assume

Codebases rarely match what their README claims (if they have one at all). Grep for signal, don't guess:

# Massive View Controller smell
grep -rlZ "class.*ViewController" --include="*.swift" --include="*.m" . \
  | xargs -0 wc -l | sort -rn | head -20

# Singleton usage density
grep -rn "\.shared\b" --include="*.swift" . | wc -l

# Delegate pattern density (common in older UIKit code)
grep -rn "Delegate" --include="*.swift" --include="*.h" . | wc -l

# MVVM signal
grep -rl "ViewModel" --include="*.swift" . | wc -l

A file count over ~400-500 lines for view controllers, high singleton density, and low ViewModel presence together indicate a Massive View Controller / MVC-in-name-only codebase regardless of what any doc claims.

Known false positives: a large view-controller file that's mostly generated or boilerplate (Interface Builder outlet dumps, generated model/mapping code) isn't the same finding as a large hand-maintained one — open the file rather than trusting the line count alone. Raw .shared count includes standard Apple singletons (UIApplication.shared, URLSession.shared, FileManager.default) that are normal API usage, not an app-specific anti-pattern — the density signal is about app-defined singletons; skim a sample of hits to see which kind dominates before citing the raw count as evidence of over-reliance. Delegate-density hits include standard UIKit protocol conformance (UITableViewDelegate, UICollectionViewDelegate) alongside any custom delegate protocols substituting for real architecture — only the latter is the smell.

3. Flag Objective-C ↔ Swift bridging points

find . -name "*-Bridging-Header.h"
find . -name "*.h" -o -name "*.m" | wc -l   # Obj-C surface area
grep -rl "@objc" --include="*.swift" . | head -20

For each bridging header found, read it and note which Obj-C types are exposed to Swift — these are the highest-risk edit points (changes on either side can silently break the other without a compiler error on both sides in mixed-language files).

4. Flag security-relevant patterns

# Hardcoded secrets/keys in source
grep -rniE '(api[_-]?key|secret|password|token)\s*=\s*"' --include="*.swift" --include="*.m" --include="*.h" .

# Credentials/tokens stored insecurely (UserDefaults instead of Keychain)
grep -rniE 'UserDefaults' --include="*.swift" . | grep -iE 'token|password|secret|credential'

# Is there a Keychain wrapper at all?
grep -rl "kSecClass\|SecItemAdd\|SecItemCopyMatching" --include="*.swift" . | wc -l

# App Transport Security weakened
grep -A5 "NSAppTransportSecurity" **/*.plist 2>/dev/null

# TLS/certificate validation bypassed
grep -rn "didReceive challenge" --include="*.swift" . -A5 | grep -i "useCredential\|performDefaultHandling"

# Sensitive data reaching logs
grep -rniE '(print|NSLog|os_log)\(.*\b(token|password|secret)\b' --include="*.swift" .

Per ios-evidence-reporting's "a pattern match is a lead, not a finding" rule, a hit on any of these is a lead to verify, not proof. The absence of any Keychain usage combined with UserDefaults calls near "token" or "password" is a strong signal worth verifying, even though neither half alone confirms insecure storage.

Known false positive: zero hits on raw kSecClass/SecItemAdd doesn't mean zero Keychain usage — a project using a third-party Keychain wrapper (e.g. KeychainAccess, KeychainSwift) calls the wrapper's API, not the raw Security framework calls this grep looks for. Check Package.resolved/Podfile.lock for a Keychain-wrapper dependency before concluding credentials aren't in the Keychain at all.

5. Map module and dependency boundaries

find . -name "Package.swift"
find . -name "Podfile"
find . -name "*.podspec"

Note whether the app is a single monolithic target, uses local SPM packages, or uses CocoaPods/Carthage — this determines how safely code can be moved or extracted later.

6. Produce the summary document

Write findings as a CLAUDE.md-ready summary with these sections:

## Codebase Summary (generated)

- **Targets:** [list from step 1]
- **Architecture actually in use:** [finding from step 2, with evidence:
  "42 view controllers over 500 lines, 180 `.shared` references,
  6 files referencing ViewModel" — not just a label]
- **Legacy Objective-C surface:** [bridging headers found, approximate
  `.m`/`.h` file count, riskiest bridging points]
- **Module structure:** [monolith / SPM packages / CocoaPods, and which]
- **Security signals:** [hits from step 4, each labeled "verify" not
  "confirmed" — e.g. "no Keychain usage found; 3 UserDefaults calls
  near 'token' in AuthManager.swift — recommend verifying token storage"]
- **Suggested entry points for new work:** [the most decoupled, best-
  tested area to build on, if one exists]

Hand this to the user as the seed for their project's own CLAUDE.md — don't silently create or overwrite one without being asked.

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.