reverse-skill
reverse-skill Features and Quickstart: Route an Authorized Security Case
A scope-first walkthrough of reverse-skill's routing rules, tool index, case initialization, evidence timeline, and report handoff.
What you will learn
- Initialize an authorized reverse-skill case
- Use routing and tool-index evidence
- Produce a replayable handoff
Before you start
- Basic Git and command-line usage
- Comfort reading a project README
You can explain the project, run its documented first step, and decide what to verify before adopting it.
Key takeaways
- Authorization and scope gates precede any active security action.
- Routing rules and tool index make methodology selection reproducible.
- Evidence hashes, timelines, and bounded handoffs turn tool output into reviewable findings.
Start with scope, not commands
reverse-skill routes APK, binary, JavaScript, PCAP, malware, CTF, firmware, API, supply-chain, and LLM-security tasks to different playbooks. Its primary ladder begins with `RULES.md`, `MASTER-ROUTING`, and `case-init`; a target should not enter an active assessment until authorization, scope, and network profile are recorded.
Use a disposable local case with a sample you own or a public CTF fixture. Keep the repository revision, case identifier, authorization note, target hash, and intended output in the run card before refreshing the tool index.
Refresh tools and choose a route
Install the repository, run the platform-specific tool-index refresh, and inspect `skills/tool-index.md`. The matrix distinguishes jadx/apktool, Frida, IDA/radare2/Ghidra, Burp/Reqable, YARA, nmap, and other tools; availability is evidence, not permission.
Run the master route against a plain-language task and record the selected R0–R44 rule, required tools, and skipped branches. If the hint is ambiguous, stop for human clarification rather than guessing a high-impact workflow.
Capture evidence as you work
Initialize `scope.md`, timeline, workitems, and a field journal. Every observation should carry a source path, command or tool, timestamp, artifact hash, and analyst note. Separate Evidence, Finding, and Path so a hypothesis cannot silently become a confirmed vulnerability.
Prefer read-only inspection first. Keep network access disabled or allow-listed until the case contract permits it, and never place credentials or real personal data in a fixture or report example.
Finish with a bounded handoff
Use case-review and docs-generator paths to produce a report that states scope, methods, evidence, limitations, remediation, and retest criteria. Run the documented routing regression and coherence checks after changing rules or manifests.
The quickstart is successful when another authorized reviewer can replay the route and locate every claim in a hashed artifact—not when a tool prints an impressive output.
Decision guide
| Criterion | Option A | Option B |
|---|---|---|
| Best when | You need predictable behavior and easy auditing | You need adaptive optimization and have reliable telemetry |
| Main risk | May leave performance on the table | Can become difficult to explain or debug |
Implementation steps
- 1
Pin the repository and create an authorized disposable case.
- 2
Refresh the platform tool index and run the master route.
- 3
Record scope, evidence, hashes, findings, and network decisions.
- 4
Run regression/coherence checks and produce a redacted handoff report.
Copy-ready example
git clone https://github.com/zhaoxuya520/reverse-skill.git
cd reverse-skill
powershell -File skills/scripts/refresh-tool-index.ps1
powershell -File skills/scripts/master-route.ps1 -Hint "analyze my APK fixture"
powershell -File skills/scripts/case-init.ps1 -CaseId demo-apkFrequently asked questions
Can I use reverse-skill against an unknown public target?
No. Use only systems, samples, or CTF targets you own or have explicit authorization to assess.
What if the route is ambiguous?
Pause for scope clarification and record the decision; do not guess a high-impact tool chain.
Sources
- reverse-skill README (captured 2026-08-31)Source checked 2026-08-31
- reverse-skill repositorySource checked 2026-08-31