ByteLandTechnology/headless-ghidra / headless-ghidra
headless-ghidra
VerifiedEntry skill for the Headless Ghidra YAML-first reverse-engineering pipeline. Use when the user asks to analyze, decompile, triage, resume, or iterate on a binary target with Ghidra/headless-ghidra. Reads artifacts/<target>/pipeline-state.yaml, routes P0–P4 phase skills, runs gate checks, and manages review pauses. Performs zero analysis work itself.
Headless Ghidra — Global Orchestrator
This skill is the workflow coordinator for the skill family. It defines the
P0–P4 sequence, dispatch rules, and artifact hand-off points.
ghidra-agent-cli remains the tool reference for command syntax and YAML
artifact semantics.
Required Shared Tool Contract
ghidra-agent-cliis the mandatory shared interface for supported workspace, metadata, Ghidra, Frida, progress, validation, and gate operations.- Phase skills must name the exact
ghidra-agent-clisubcommands they use. - Lower-level shell scripts and Java helpers are backend details. They must not replace the CLI as the primary interface when the CLI already supports the action.
- All workflow artifacts must live under
artifacts/<target-id>/. - YAML artifacts must be created, updated, and validated by
ghidra-agent-cli. - The CLI must not automatically create git commits.
- Gate transitions require relevant artifacts to exist on disk and be ready for user review.
- All Ghidra project operations must go through
ghidra-agent-cli. If the CLI lacks a required capability, pause and ask the user before creating or running a new Ghidra script.
Pipeline
P0 Intake → P1 Baseline+Runtime → P2 Third-Party → [P3 Metadata Enrichment → P4 Function Substitution]*
| Phase | Skill | Purpose | Primary outputs |
|---|---|---|---|
| P0 | headless-ghidra-intake | Initialize target workspace, discover prerequisites, and define scope | pipeline-state.yaml, scope.yaml, targets/<id>/ghidra-projects/ |
| P1 | headless-ghidra-baseline | Run Ghidra import/analysis, export baseline YAML, and prepare runtime observations | baseline/*.yaml, runtime/run-manifest.yaml, runtime/run-records/*.yaml, runtime/hotpaths/call-chain.yaml |
| P2 | headless-ghidra-evidence | Identify and record third-party libraries and pristine sources | third-party/identified.yaml, third-party/pristine/<library>@<version>/, third-party/compat/<library>@<version>/ |
| P3 | headless-ghidra-discovery | Enrich names, signatures, types, constants, strings, and selected hotpath metadata | metadata/*.yaml, metadata/apply-records/ |
| P4 | headless-ghidra-batch-decompile | Substitute selected functions through metadata application and Ghidra decompilation | substitution/next-batch.yaml, substitution/functions/<fn_id>/ |
Shared Artifact Contract
All phases work inside this active workspace layout:
targets/<target-id>/ghidra-projects/
artifacts/<target-id>/
├── pipeline-state.yaml
├── scope.yaml
├── intake/
├── baseline/
├── runtime/
│ ├── project/
│ ├── fixtures/
│ ├── run-manifest.yaml
│ ├── run-records/
│ └── hotpaths/call-chain.yaml
├── third-party/
│ ├── identified.yaml
│ ├── pristine/<library>@<version>/
│ └── compat/<library>@<version>/
├── metadata/
│ ├── renames.yaml
│ ├── signatures.yaml
│ ├── types.yaml
│ ├── constants.yaml
│ ├── strings.yaml
│ └── apply-records/
├── substitution/
│ ├── template/
│ ├── next-batch.yaml
│ └── functions/<fn_id>/
└── gates/
The orchestrator treats pipeline-state.yaml as the current target-level state
record and relies on the phase-owned YAML artifacts above for hand-offs.
Orchestrator Responsibilities
- Detect or resume the active target.
- Read
artifacts/<target-id>/pipeline-state.yaml. - Dispatch the correct phase skill for the current stage.
- Run
ghidra-agent-cli gate check --phase ...at each transition. - Advance phase state only after the gate passes.
- Handle user dialogs such as resume/restart, optional Frida supplementation, batch confirmation, divergence review, and completion.
Gate Policy
- P0–P4 are the only primary pipeline transitions.
ghidra-agent-cli gate checkis the authoritative gate validation for all pipeline phases (P0–P4). The legacygate-check.shhas been removed.
Required ghidra-agent-cli Commands
ghidra-agent-cli context useghidra-agent-cli context showghidra-agent-cli context clearghidra-agent-cli workspace state showghidra-agent-cli workspace state set-phaseghidra-agent-cli gate checkghidra-agent-cli validateghidra-agent-cli progress compute-next-batchghidra-agent-cli progress show
Strict Prohibitions
- Must not execute analysis work itself.
- Must not edit baseline, evidence, decompilation, or verification artifacts directly except for explicit state updates it owns.
- Must not bypass
ghidra-agent-clifor supported state, progress, context, validation, or gate operations. - Must not accept alternate decompilation backends in place of Ghidra.
- Must not create git commits automatically.
- Must not create or run new Ghidra scripts when the CLI lacks a capability; pause and ask the user first.
Next Skill Routing
- P0 complete →
headless-ghidra-baseline - P1 complete →
headless-ghidra-evidence - P2 complete →
headless-ghidra-discovery - P3 complete →
headless-ghidra-batch-decompile - P4 complete for all selected functions → either loop back to P3 or finish
Independent Skills
The following skills operate outside the P0–P4 pipeline and can be invoked directly:
| Skill | Purpose | Invocation |
|---|---|---|
headless-ghidra-analyze-function | Thorough single-function analysis following the strict five-step recovery order: types → constants → vtables → function identity → decompilation. Use when you need complete analysis of one specific function with full type/constant/vtable context before decompilation results are interpreted. | Invoked when user asks to "analyze this function thoroughly", "decompile function at 0x... with full context", or requests complete per-function analysis. |
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| writing-skillsobra/superpowersUse when creating new skills, editing existing skills, or verifying skills work before deployment | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| xlsxanthropics/skillsUse this skill any time a spreadsheet file is the primary input or output. This means any task where the user wants to: open, read, edit, or fix an existing .xlsx, .xlsm, .xltx, .csv, or .tsv file (e.g., adding columns, computing formulas, formatting, charting, cleaning messy data); create a new spreadsheet from scratch or from other data sources; or convert between tabular file formats. Trigger especially when the user references a spreadsheet file by name or path — even casually (like "the xlsx in my downloads") — and wants something done to it or produced from it. Also trigger for cleaning or restructuring messy tabular data files (malformed rows, misplaced headers, junk data) into proper spreadsheets. The deliverable must be a spreadsheet file. Do NOT trigger when the primary deliverable is a Word document, HTML report, standalone Python script, database pipeline, or Google Sheets API integration, even if tabular data is involved. | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |
| shipping-and-launchaddyosmani/agent-skillsPrepares production launches. Use when preparing to deploy to production, or when asking what needs to be in place before shipping. Use when you need a pre-launch checklist, when setting up monitoring, when planning a staged rollout, or when you need a rollback strategy. | Stars: 119k | Updated | Claude CodeCodex | Verified | Install |
| childhoodthedotmack/claude-memBabysit a shipped release after merge, publication, or deployment. Verify published packages, production behavior, installed plugins, and late CI or user reports until the observation window is clean; use babysit for PR readiness before release. | Stars: 118k | Updated | Claude CodeCodex | Verified | Install |