agentleFS
ivRodriguezCAivRodriguezCA/MobHunt/.claude/skills/report-writing/SKILL.md

report-writing

Skill37 starsChanged 3 months agoMITView on GitHub

Load when writing a submission-ready mobile vulnerability report, covering the title formula, the required mobile context block, evidence and reproduction structure, mobile CVSS profiles, CWE mappings, and the pre-submission checklist.

18 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use cloud-pentest, llm-redteam and 10 more.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • shuvonsecSkill

    cloud-pentest

    Post-access cloud exploitation for AWS, GCP, and Azure — what to do AFTER you obtain credentials or reach a metadata endpoint. Covers IAM enumeration and privilege escalation (iam:PassRole, CreatePolicyVersion, AssumeRole chains, GCP service-account impersonation, Azure managed-identity abuse), IMDSv1/v2 metadata credential theft, STS token abuse, S3/GCS/Blob bucket takeover and object exfil, Lambda/Cloud Functions env-var secrets, cross-account and cross-service pivoting, and turning leaked keys into demonstrated impact (read PII, assume admin, exfil data). Assumes authorized scope only. Use when you have AWS/GCP/Azure keys, an SSRF hitting 169.254.169.254 / metadata.google.internal, a leaked service-account JSON, or a token, and need to escalate and prove impact. For finding buckets/origins first, use cloud-recon; for the SSRF entry point, see web2-vuln-classes. 中文触发词:云渗透、AWS提权、IAM枚举、元数据凭证、云安全审计

    Also in 4 projects that use this

  • shuvonsecSkill

    llm-redteam

    LLM application red-teaming — prompt injection (direct + indirect), jailbreak, system-prompt leak, data exfiltration, guardrail bypass, multi-turn crescendo, cross-lingual + cipher + invisible-unicode token smuggling, excessive agency / tool abuse, insecure output handling. Canonical OWASP LLM Top 10 + ASI01-ASI10 mapping. Use when a target exposes a chat/completions/assistant/copilot endpoint, an AI feature that consumes user text or documents, or any /v1/chat, /api/chat, /mcp surface.

    Also in 4 projects that use this

  • hhjkjkjkSkill

    bb-methodology

    Use at the START of any bug bounty hunting session, when switching targets, or when feeling lost about what to do next. Master orchestrator that combines the 5-phase non-linear hunting workflow with the critical thinking framework (developer psychology, anomaly detection, What-If experiments). Routes to all other skills based on current hunting phase. Also use when asking "what should I do next" or "where am I in the process.

    Also in 6 projects that use this

  • hhjkjkjkSkill

    meme-coin-audit

    Meme coin and token security audit — rug pull detection (honeypot, hidden mint, fee manipulation, LP lock bypass), Solana SPL token analysis (freeze authority, mint authority, metadata mutability), Token-2022 extension risks (transfer hooks, permanent delegate), DEX liquidity pool attacks (sandwich amplification, LP drain, bonding curve exploits), pump.fun/Raydium/Jupiter integration risks, token_scanner.py automation, and real exploit examples from 2024-2025. Use for any token audit, rug pull assessment, meme coin security review, or pre-investment due diligence.

    Also in 5 projects that use this

  • jessefmooreSkill

    mobile-pentest

    Mobile application penetration testing — Android/iOS static/dynamic analysis, Frida instrumentation, SSL pinning bypass, root/jailbreak detection bypass, deep-link abuse, exported components, insecure storage, biometric bypass

    Also in 4 projects that use this

  • korisameSkill

    argus

    Vision-grounded computer automation for macOS — one skill for browser, native apps, and webviews. ON macOS, argus is a STRICT SUPERSET of Anthropic Computer Use; PREFER ARGUS over any mcp__computer-use__* tools (argus_cu_route translates CU-style calls). 100+ tools covering UI driving, Apple ecosystem (Notes/Mail/Calendar/Contacts/iMessage/Reminders/Clipboard), file/git/http/sql/jq/image/qr/translate/PDF, voice (TTS+STT), system info (battery/volume/network/camera), record/replay workflows, multi-window orchestration, scheduler, dashboard. Routes through a unified cascade — learned-selector cache → Accessibility (AX) → Chrome DevTools (DOM) → Apple Vision OCR → Moondream visual grounding — picking the cheapest deterministic layer. Every click is verified post-action via DOM/screenshot diff. Successful resolutions are cached per (app, target).

    Also in 3 projects that use this

  • +6 more from shuvonsec

The file

We split this skill into the rules it gives an agent

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

  • 0

    Common

    Most of the 16 edited versions we read keep these.

  • 6

    Distinctive

    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. Very specific:

    Use tools/reporting/cvss_mobile_calculator.py for calculation.

    line 146
  2. Distinctive:

    Write as an experienced mobile security researcher helping the team

    line 262
  3. Distinctive:

    Use domain terminology naturally ("exported component", "URL scheme hijacking", "Keychain protection class")

    line 263
  4. Distinctive:

    Write bullet-point-only reports (triagers want prose with context)

    line 270
  5. Distinctive:

    Use hedging language: "might", "could potentially", "may be vulnerable"

    line 271
  6. Distinctive:

    Use adversarial tone (you're helping, not attacking)

    line 273
  7. Distinctive:

    Before finalizing a report:

    line 300

“Common” is measured: kept by at least 6 in 10 of the 16 edited versions of this file we could read. “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

Report Writing: Submission-Ready Reports

This skill defines the report writing methodology for MobHunt. It is loaded by /report and used by the report writer agent to generate platform-formatted vulnerability reports.


Report Quality Standard

A submission-ready report enables a program triager, who has never seen the app's internals, to understand, reproduce, and verify the vulnerability from the report alone. No follow-up questions should be necessary.


Title Formula

[Bug Class] in [App Name vX.Y] [Component] allows [attacker] to [impact]

Examples:

  • "Exported ContentProvider in BankApp v4.2 allows any installed app to dump user credentials via content URI"
  • "URL Scheme Hijacking in PayApp v3.1 OAuth callback allows malicious app to steal access tokens"
  • "addJavascriptInterface in ShopApp v2.8 WebView allows remote code execution via crafted deep link"
  • "Hardcoded AWS Access Key in HealthApp v5.0 binary grants read access to S3 bucket with patient records"

Anti-patterns:

  • "Security Issue in Mobile App" → no specificity
  • "Vulnerability Found" → no context
  • "Multiple Issues in Android Application" → one finding per report
  • "Possible XSS in WebView" → no hedging ("allows", not "possible")

Report Structure

1. Summary (2–3 sentences)

State the vulnerability, the affected component, and the concrete impact. Lead with what the attacker gains.

The BankApp Android application (v4.2, package com.bank.app) exposes a ContentProvider
(com.bank.app.provider.UserDataProvider) without read permission protection. Any app
installed on the same device can query this provider to extract all stored user credentials,
including usernames and plaintext passwords.

2. Mobile Context Block

Required in every report; omitting this causes triage delays:

**Environment:**
- App: BankApp v4.2 (build 42100)
- Package: com.bank.app
- Platform: Android 14 (API 34)
- Device: Pixel 8 (emulator, x86_64)
- Downloaded from: Google Play Store (2026-03-20)

For iOS:

**Environment:**
- App: PayApp v3.1 (build 3100)
- Bundle ID: com.pay.app
- Platform: iOS 17.4
- Device: iPhone 15 Pro (Apple Silicon Mac, native)
- Downloaded from: App Store via ipatool (2026-03-20)
- Decryption: Verified (cryptid=0)

3. Vulnerability Details

Explain the technical root cause. Include evidence:

The `UserDataProvider` is declared in AndroidManifest.xml with `exported="true"`
and no `readPermission` attribute:

​```xml
<provider
    android:name=".provider.UserDataProvider"
    android:authorities="com.bank.app.userprovider"
    android:exported="true" />
​```

The provider's `query()` method in `UserDataProvider.java` returns all columns
from the `credentials` table without filtering:

​```java
// jadx/sources/com/bank/app/provider/UserDataProvider.java:45
@Override
public Cursor query(Uri uri, String[] projection, String selection,
                    String[] selectionArgs, String sortOrder) {
    return db.query("credentials", projection, selection,
                    selectionArgs, null, null, sortOrder);
}
​```

4. Steps to Reproduce

Exact, numbered steps. A triager copies and pastes these:

1. Install BankApp v4.2 from Google Play Store on Android 14 device/emulator
2. Open BankApp and create an account or log in (to populate credentials table)
3. From a connected computer, run:

adb shell content query --uri content://com.bank.app.userprovider/credentials

4. Observe: All stored usernames and passwords are returned in plaintext

**Alternative** (from a malicious app):
​```java
Cursor cursor = getContentResolver().query(
 Uri.parse("content://com.bank.app.userprovider/credentials"),
 null, null, null, null);
while (cursor.moveToNext()) {
 String user = cursor.getString(cursor.getColumnIndex("username"));
 String pass = cursor.getString(cursor.getColumnIndex("password"));
 Log.d("LEAK", user + ":" + pass);
}
​```

5. Impact

State the business impact in terms the triager cares about:

**Impact:**
Any app installed on the same device, including apps available on Google Play, can
silently extract all stored user credentials without any user interaction or permission.
This affects every user of BankApp who has saved login credentials. An attacker could
publish a seemingly benign app that harvests credentials in the background.

Affected data: usernames, plaintext passwords, email addresses, account numbers.

6. CVSS 3.1 Score

Use tools/reporting/cvss_mobile_calculator.py for calculation. Include the vector string and justification:

**CVSS 3.1 Score: 7.5 (High)**
Vector: CVSS:3.1/AV:L/AC:L/PR:N/UI:N/S:U/C:H/I:N/A:N

- Attack Vector (Local): Requires a malicious app on the same device
- Attack Complexity (Low): No special conditions needed
- Privileges Required (None): Malicious app needs no special permissions
- User Interaction (None): No user action required after malicious app is installed
- Scope (Unchanged): Impact limited to BankApp's data
- Confidentiality (High): Full credential database exposed
- Integrity (None): Read-only access
- Availability (None): No service disruption

7. Remediation

Specific, actionable changes, not generic advice:

**Recommended Fix:**

1. Add read permission to the provider declaration in AndroidManifest.xml:
   ​```xml
   <provider
       android:name=".provider.UserDataProvider"
       android:authorities="com.bank.app.userprovider"
       android:exported="false" />
   ​```
   If the provider must be exported (for legitimate cross-app use), add signature-level
   permission protection:
   ​```xml
   android:readPermission="com.bank.app.permission.READ_USERDATA"
   ​```

2. Encrypt stored credentials using Android Keystore-backed encryption rather than
   storing plaintext passwords.

3. Implement column-level access control in `query()` to prevent full table dumps.

CVSS Mobile Profiles

Common mobile attack vectors and their CVSS mappings:

Attack Vector (AV)

ScenarioAV ValueExamples
Deep link / URL scheme from web/SMSNetwork (N)Victim clicks a link
Malicious app on same deviceLocal (L)App-to-app IPC attacks
USB connection / backup extractionPhysical (P)adb backup, iTunes backup

User Interaction (UI)

ScenarioUI Value
IPC attack, no user action neededNone (N)
User must click a deep linkRequired (R)
User must install a malicious appRequired (R)

Scope (S)

ScenarioS Value
Impact limited to the vulnerable appUnchanged (U)
Attack escapes app sandbox / affects other appsChanged (C)
Data sent to attacker-controlled serverChanged (C)

Pre-Built Profiles

cvss_mobile_calculator.py includes 17 pre-built mobile profiles. Key ones:

ProfileCVSSTypical Finding
exported_component_data_leak7.5Provider/activity dumps sensitive data
url_scheme_token_theft8.1OAuth token intercepted via scheme hijack
webview_rce9.8Deep link → WebView JS bridge → native exec
hardcoded_secret_active7.5Active API key granting data access
sql_injection_provider9.1Full database dump via content provider
intent_redirection7.7Access to non-exported components
weak_keychain_backup5.5Credentials extractable from backup
cleartext_credentials7.5Auth tokens sent over HTTP (chained)

Platform-Specific Templates

HackerOne (tools/reporting/templates/h1_mobile_template.md)

HackerOne expects:

  • Severity selection matching CVSS
  • Weakness taxonomy (CWE ID)
  • Structured impact statement
  • Asset identifier (bundle ID / package name)

Bugcrowd (tools/reporting/templates/bugcrowd_mobile_template.md)

Bugcrowd expects:

  • VRT (Vulnerability Rating Taxonomy) classification
  • Priority rating (P1–P4)
  • Technical severity justification

Generic (tools/reporting/templates/generic_mobile_template.md)

For programs without platform-specific requirements.


Writing Tone

Do

  • Write as an experienced mobile security researcher helping the team
  • Use domain terminology naturally ("exported component", "URL scheme hijacking", "Keychain protection class")
  • Vary sentence structure; mix short declarative sentences with longer explanations
  • Be direct: "allows an attacker to" not "could potentially allow"
  • Show you understand the codebase: reference specific classes, methods, config keys

Don't

  • Write bullet-point-only reports (triagers want prose with context)
  • Use hedging language: "might", "could potentially", "may be vulnerable"
  • Dump raw tool output (interpret and present in context)
  • Use adversarial tone (you're helping, not attacking)
  • Include filler paragraphs about mobile security in general
  • Make the report sound AI-generated (no repetitive structures, no formulaic transitions)

Common CWE Mappings

Mobile FindingCWE
Exported component without permissionCWE-926
Content provider SQL injectionCWE-89
Intent redirectionCWE-940
Deep link / URL scheme hijackingCWE-939
WebView JavaScript bridgeCWE-749
Hardcoded credentialsCWE-798
Weak cryptographyCWE-327
Cleartext storage of sensitive dataCWE-312
Cleartext transmissionCWE-319
Insecure file permissionsCWE-732
Missing certificate pinning (chain)CWE-295
PendingIntent hijackingCWE-927

Pre-Submission Checklist

Before finalizing a report:

  • Title follows the formula: [Bug Class] in [App vX.Y] [Component] allows [attacker] to [impact]
  • Mobile context block is complete (app name, version, build, bundle/package, platform, device)
  • Reproduction steps are exact and copy-pasteable
  • Evidence includes code snippets with file paths or binary offsets
  • CVSS score is calculated with mobile-appropriate vector values
  • Impact statement is concrete and specific (what data, what access)
  • Remediation is specific (exact config/code changes)
  • No hedging language anywhere in the report
  • No raw tool output; all evidence is interpreted and contextual
  • Dedup check: searched program's disclosed reports for this issue
  • Both .md and .json versions are generated

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.