메뉴
Canonical fuzzing workflow for confirmed/proven findings. Coordinates plan → author harnesses (fuzz-harness-author agent) → replay → triage → promote, and advances a finding to proven only on empirical crash/sanitizer evidence. Prefer this over the lower-level /fuzz-* stage commands.
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
SOC 직업 분류 기준
Interprocedural, hypothesis-driven hunt — finds the multi-file source→sink bugs that pattern-gating and same-file taint miss. Ranks trace anchors (entry points + dangerous sinks), then the deep-hunter agent walks each flow ACROSS files (forward/backward call-graph CLIs) over multiple rounds, confirming propagation and defeating guards, and promotes confirmed cross-file flows into findings.json with the path stored as evidenceGraph. Token-expensive, budget-bounded. Part of the HUNT phase (run via /sweep --deep, or ask for it); leads flow to /verify (panel).
Whole-file deep reader — finds bugs by READING risk-ranked files in full and reasoning from first principles, not by grepping for known patterns. Catches the long tail (project-specific wrappers, business logic, cross-function flaws) that pattern-gated producers structurally miss. Token-expensive; budget-bounded and risk-ranked. Promotes leads into .kuzushi/findings.json (source "deep-scan").
IRIS-style source→sink taint hunt. Ranks a typed CWE catalog for the repo, then drives subagents to label dangerous sinks, label sources of user input, run Joern/CodeQL dataflow queries (or same-file linking) to connect them, and triage each flow as finding/candidate/rejected. Promotes verdicts into .kuzushi/findings.json. Benefits from a prebuilt CodeQL DB / Joern CPG but degrades gracefully.
Adversarial per-threat review (Carlini doctrine). For each threat in .kuzushi/threat-model.json, state attacker capabilities, walk source→sink, attempt to bypass every guard, and assign a verdict with evidence. Promotes findings into .kuzushi/findings.json. Requires /threat-model first.
Split the attack surface into parallel-discovery partitions so fan-out hunters explore different subsystems instead of converging on the same shallow bug. Deterministic grouping of /x-ray entry points by component → .kuzushi/partitions.json, which a hunt coordinator hands to one subagent per partition. Run /x-ray first.
Empirical proof-of-concept for the PoC-ready findings. For each finding /verify marked confirmed-exploitable or inconclusive, synthesize a minimal harness that triggers the bug; a host script then runs it in a sandbox (Docker --network none when present, else a gated local run) and classifies the crash into a proof verdict. Attaches a poc block onto each finding. Requires /verify first.
| name | fuzz |
| description | Canonical fuzzing workflow for confirmed/proven findings. Coordinates plan → author harnesses (fuzz-harness-author agent) → replay → triage → promote, and advances a finding to proven only on empirical crash/sanitizer evidence. Prefer this over the lower-level /fuzz-* stage commands. |
| user-invocable | false |
Use /fuzz as the main workflow — you are the coordinator: run the deterministic stages and
spawn the harness-authoring subagent between them. The lower-level /fuzz-init, /fuzz-run,
/fuzz-triage, /fuzz-minimize, and /fuzz-promote commands are replay/debug stages behind this
surface.
node "${CLAUDE_PLUGIN_ROOT}/scripts/cmd/fuzz.mjs" --target "<repo root>" --stage plan
If it returns no-seeds, tell the user to run /verify first and stop.fuzz-harness-author agent with the Task tool, passing it the target directory and
the .kuzushi/fuzz/fuzz-plan.json path. It writes a real harness into each candidate's
harnessDir, seeds the corpus, sets a concrete offline runCommand + expectedSignal (using the
recommended engine + semanticOracle controls), and writes the updated plan back in place. Wait
for it to finish.node "${CLAUDE_PLUGIN_ROOT}/scripts/cmd/fuzz.mjs" --target "<repo root>" --stage replay
Add --trust-local only if Docker is unavailable and the user explicitly
accepts local execution.proofVerdict:"exploited" promotes.