Ejecuta cualquier Skill en Manus
con un clic

Ejecuta cualquier Skill en Manus con un clic

deep-context

Estrellas0

Forks0

Actualizado29 de mayo de 2026 a las 02:51

Deep system-understanding pass before threat modeling. The context-analyst agent reads the code (line-by-line where it matters) and builds a grounded model — modules, entry points, actors, trust boundaries, data stores, system invariants — with file:line evidence, written to .kuzushi/deep-context.json. Context only — it never finds vulns, fixes, or severities. Best run after x-ray, before /threat-model.

Instalación

Instalar con Codex o Claude Copia este prompt, pégalo en Codex, Claude u otro asistente, y deja que revise la página de la skill y la instale por ti.

Ejecutar en Manus

Fuente

allsmog

allsmog/kuzushi-security-plugin

Abrir repositorio de GitHub Ver repositorios del creador

Descarga

Ejecutar en Manus

Ocupaciones relacionadasSOC

Basado en la clasificación ocupacional SOC

Analistas de seguridad de la informaciónOcupaciones informáticas y matemáticas·SOC 15-1212

SKILL.md

readonly

name	deep-context
description	Deep system-understanding pass before threat modeling. The context-analyst agent reads the code (line-by-line where it matters) and builds a grounded model — modules, entry points, actors, trust boundaries, data stores, system invariants — with file:line evidence, written to .kuzushi/deep-context.json. Context only — it never finds vulns, fixes, or severities. Best run after x-ray, before /threat-model.
context	fork
agent	context-analyst
user-invocable	false

Deep context

Build a grounded understanding of the system before hunting bugs (shallow context is where missed bugs and false positives come from).

Run node "${CLAUDE_PLUGIN_ROOT}/scripts/cmd/deep-context-prepare.mjs" --target "<repo root>". Read the prep's prepPath (inventory + x-ray entry points as leads).
Work the three phases — orientation (modules / entry points / actors / data stores) → ultra-granular analysis of the security-relevant components (read line-by-line, trace cross-function flows, record the invariants the code assumes) → a global data-flow + trust model — grounding every claim in a file you read (tree_sitter:*, LSP, Grep). Record anything suspicious as an open question, never a finding.
Write the { systemOverview, modules, entryPoints, actors, trustBoundaries, dataStores, invariants, openQuestions } bundle to the prep's draftPath, then run the assembleCommand — it enforces the context-only boundary and persists .kuzushi/deep-context.json.
Report the overview, the counts, the key invariants, and the open questions. /threat-model will build on it.

When NOT to use

When a quick inventory is enough — that's the SessionStart context / /threat-model's own scope; this is the slower, deeper reasoning pass.
To find or fix bugs — strictly out of scope here; use /threat-hunt, /taint-analysis, /systems-hunt, and /verify.

Rationalizations to Reject

"The framework implies the design." → Read the code; assumed designs miss real trust boundaries.
"That looks like a bug — flag it." → Record an openQuestion; the hunts adjudicate, not this stage.
"Skip the confusing module." → That's often where bugs hide; if you can't resolve it, say so in openQuestions rather than glossing over it.

Más de este repositorio

mismo repositorio

deep-hunt

allsmog/kuzushi-security-plugin

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).

2026-06-110

deep-scan

allsmog/kuzushi-security-plugin

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").

2026-06-110

taint-analysis

allsmog/kuzushi-security-plugin

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.

2026-06-110

threat-hunt

allsmog/kuzushi-security-plugin

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.

2026-06-110

partition

allsmog/kuzushi-security-plugin

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.

2026-06-100

poc

allsmog/kuzushi-security-plugin

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.

2026-06-100

name	deep-context
description	Deep system-understanding pass before threat modeling. The context-analyst agent reads the code (line-by-line where it matters) and builds a grounded model — modules, entry points, actors, trust boundaries, data stores, system invariants — with file:line evidence, written to .kuzushi/deep-context.json. Context only — it never finds vulns, fixes, or severities. Best run after x-ray, before /threat-model.
context	fork
agent	context-analyst
user-invocable	false

Deep context

Build a grounded understanding of the system before hunting bugs (shallow context is where missed bugs and false positives come from).

Run node "${CLAUDE_PLUGIN_ROOT}/scripts/cmd/deep-context-prepare.mjs" --target "<repo root>". Read the prep's prepPath (inventory + x-ray entry points as leads).
Work the three phases — orientation (modules / entry points / actors / data stores) → ultra-granular analysis of the security-relevant components (read line-by-line, trace cross-function flows, record the invariants the code assumes) → a global data-flow + trust model — grounding every claim in a file you read (tree_sitter:*, LSP, Grep). Record anything suspicious as an open question, never a finding.
Write the { systemOverview, modules, entryPoints, actors, trustBoundaries, dataStores, invariants, openQuestions } bundle to the prep's draftPath, then run the assembleCommand — it enforces the context-only boundary and persists .kuzushi/deep-context.json.
Report the overview, the counts, the key invariants, and the open questions. /threat-model will build on it.

When NOT to use

When a quick inventory is enough — that's the SessionStart context / /threat-model's own scope; this is the slower, deeper reasoning pass.
To find or fix bugs — strictly out of scope here; use /threat-hunt, /taint-analysis, /systems-hunt, and /verify.

Rationalizations to Reject

"The framework implies the design." → Read the code; assumed designs miss real trust boundaries.
"That looks like a bug — flag it." → Record an openQuestion; the hunts adjudicate, not this stage.
"Skip the confusing module." → That's often where bugs hide; if you can't resolve it, say so in openQuestions rather than glossing over it.