agentleFS
zyphr-labszyphr-labs/bounty-harness/.claude/skills/planner/SKILL.md

planner

Skill1 starsChanged 3 months agoMITView on GitHub

This replaces "did we run everything?" with "here is a measurable plan covering X% of relevant surfaces." This is a thin human-facing router. Use skill.yaml as the source of truth for exact commands, inputs, outputs, and workflow chaining. Prerequisite: The planner requires a domain profile. If the target has no prior recon, run the recon skill first, then the domain-model skill to generate a domain profile, then use the planner.

35 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use executor, reviewer and 6 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.
  • MooseDesign1Skill

    executor

    Main executor agent. Triggered by /executor, or natural language such as: "build this feature", "execute the feature", "start building", "map the codebase", "scan the code", "resume building", "continue the build", "dev status", "where are we on development", "what's built". Do NOT trigger for /planner, /plan:project, /plan:feature, /plan:status, /product-designer, /pd:*, /ds-designer, /ds-*.

    Also in 3 projects that use this

  • MicrosoftSkill

    reviewer

    ASP.NET Core ハンドブックの記事をレビューする。Use when: 記事をレビューする、ドキュメントを校正する、コードレビューする、PRをレビューする。技術ドキュメントのレビュー、品質チェック、校正。

    Also in 5 projects that use this

  • JohnJBorenSkill

    debugger

    Systematic debugging workflows, root cause analysis, and troubleshooting patterns

    Also in 3 projects that use this

  • JetBrainsSkill

    review-plan

    Manually re-run the autonomous plan review (Phase 2 — consistency + structural). The same review runs automatically as the first phase of /execute-tracks; this command is for re-runs after inline replanning or whenever the plan needs explicit re-validation. TRIGGER when: user asks to re-validate the plan after inline replanning; user requests explicit plan review outside /execute-tracks. SKIP: /execute-tracks already running State 0 in this session (autonomous Phase 2 path will fire) — do not double-run.

    Also in 3 projects that use this

  • tesla0225Skill

    orchestrator

    Issue/要件を受けたら分解→担当割当→実装→検証→統合を回す手順。未整備なら /repo-init でbootstrap。

    Also in 4 projects that use this

  • Harut8Skill

    architect

    Also in 3 projects that use this

The file

We split this skill into the rules it gives an agent

4 rules. 0 are common, 0 are distinctive, and 4 are very specific to this project.

  • 0

    Common

    Most of the 31 edited versions we read keep these.

  • 0

    Distinctive

    This author's own choices. Worth a look.

  • 4

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Start with generate-plan for the full plan.

    line 40
  2. Very specific:

    Use generate-plan-safe if you only want safe, non-destructive items.

    line 41
  3. Very specific:

    Follow with visualize-plan and validate-plan to review before execution.

    line 42
  4. Very specific:

    After execution, use the plan-vs-results runbook to feed findings back into the coverage matrix.

    line 43

“Common” is measured: kept by at least 6 in 10 of the 31 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

Planner

Overview

Domain-driven ranked test plan generator. Consumes the Domain Model (archetypes + surfaces), the Asset Graph (recon data), the Technique KB catalog, and the Coverage Ledger. Outputs a prioritized list of workflows to execute — ranked by business impact, surface prevalence, vulnerability severity, signal quality, coverage gaps, and tool availability.

This replaces "did we run everything?" with "here is a measurable plan covering X% of relevant surfaces."

This is a thin human-facing router. Use skill.yaml as the source of truth for exact commands, inputs, outputs, and workflow chaining.

Quick Reference

  • Skill: planner
  • Severity range: info
  • Required tools: python3, jq, yq
  • Expected input files: Domain profile JSON (from domain-model/), Technique YAML catalog (from technique-kb/), optional coverage matrix YAML or JSON
  • Scope check: The planner itself does not execute intrusive tests — it plans them. Ensure scope authorization exists before executing any plan items.
  • Secret handling: Any auth headers, session tokens, cookies, or bearer credentials referenced in domain profiles or plan items are local-only and must be redacted before sharing; sanitize plan artifacts and never commit secrets (covered by gitignore).

When to Use

Trigger phrases:

  • "create a test plan", "generate a plan", "ranked plan", "what should I test", "prioritize my testing", "plan my engagement", "coverage-based plan"
  • "plan", "planning", "test strategy", "engagement plan", "plan vs results"

Prerequisite: The planner requires a domain profile. If the target has no prior recon, run the recon skill first, then the domain-model skill to generate a domain profile, then use the planner.

Decision Tree

User provides target + wants a plan
          │
          ├── Domain profile exists? ── No ──> Run recon → domain-model → retry
          │
          ├── Technique catalog populated? ── No ──> Warn user, plan will be empty
          │
          ├── Coverage matrix available? ── No ──> Plan treats all techniques as gaps
          │
          └── Generate plan ──> Review ──> Execute ──> Plan vs Results

Workflow Selection

  • Start with generate-plan for the full plan.
  • Use generate-plan-safe if you only want safe, non-destructive items.
  • Follow with visualize-plan and validate-plan to review before execution.
  • After execution, use the plan-vs-results runbook to feed findings back into the coverage matrix.

Available Workflows

WorkflowPurposeScript pathsPrimary outputs
generate-planGenerate full ranked test plan from domain profile + techniques + coverage.claude/skills/planner/scripts/generate_plan.pyPlan JSON file
generate-plan-safeGenerate plan excluding intrusive and destructive techniques.claude/skills/planner/scripts/generate_plan.pyPlan JSON file
visualize-planConvert plan JSON to readable markdown or HTML.claude/skills/planner/scripts/plan_visualizer.pyMarkdown or HTML file
validate-planValidate plan JSON against plan_schema.yaml for structural correctness.claude/skills/planner/scripts/plan_validator.pyValidation report (stdout)
import-workflowImport portable open-kritt-workflow v2 JSON into a playbook YAMLtools/import_workflow.pyPlaybook YAML file
apply-model-overridesApply per-depth model overrides to a plan (Kritt-style).claude/skills/planner/scripts/model_overrides.pyPlan with model_override per item

Runbooks

RunbookPurposePath
Generate PlanStep-by-step plan generation.claude/skills/planner/runbooks/generate-plan.md
Review PlanReview plan for safety, scope, completeness.claude/skills/planner/runbooks/review-plan.md
Execute PlanExecute plan items, collect evidence, record results.claude/skills/planner/runbooks/execute-plan.md
Plan vs ResultsCompare plan to actual results, update coverage matrix.claude/skills/planner/runbooks/plan-vs-results.md

Scoring Model

Each technique is scored on 6 dimensions and weighted to produce a final score [0,1]:

DimensionWeightWhat it measures
Business Impact0.30Archetype severity weight × technique severity mapping
Surface Prevalence0.20Match between technique surfaces and target's detected surfaces
Vulnerability Severity0.20Direct mapping from technique severity field
Detection Signal Quality0.15Number and specificity of positive/negative expected signals
Coverage Gap Urgency0.10Whether the technique fills a prior gap (1.0 = not covered, 0.0 = covered)
Tool Availability0.05Fraction of required tools found in PATH

Priority tiers: critical (≥0.80), high (≥0.60), medium (≥0.35), low (<0.35)

Technique YAML Format (expected in technique-kb/techniques/)

id: "xss-reflected-get"
name: "Human-readable technique name"
category: "xss"
severity: "critical|high|medium|low|info"
description: "What this technique detects"
standards:
  wstg: ["WSTG-INPV-01"]
  asvs: []
  api_top10: []
  vrt: []
  cwe: ["CWE-79"]
applies_to:
  archetypes: ["all"]
  surfaces: ["all-parameterized-endpoints"]
requires:
  auth: "none"
  inputs: ["parameterized URLs"]
  tools: ["curl", "dalfox"]
signals:
  positive:
    - "Payload echoed in response without sanitization"
  negative:
    - "Content-Type prevents script execution"
safety:
  intrusive: false
  data_modifying: false
  rate_limited: false
  requires_approval: false
evidence:
  required:
    - "raw_request"
    - "raw_response"
    - "screenshot"
workflow_mapping:
  skill: "xss"
  workflow: "reflected-param-test"

Evidence Required

The planner itself does not produce evidence — it produces a plan. Each plan item specifies what evidence to collect during execution:

  • raw_request — full HTTP request with headers
  • raw_response — full HTTP response with headers and body
  • screenshot — visual evidence of the finding
  • poc_script — reproducible proof of concept
  • tool_output — raw tool output logs
  • timestamp — ISO 8601 execution time

Output Files

FileContents
plan.jsonFull ranked plan in plan_schema format
plan.md / plan.htmlHuman-readable visualizations
coverage.jsonUpdated coverage matrix (after plan-vs-results)

References

  • Source of truth: skill.yaml
  • Scoring rules: ranking_rules.yaml
  • Plan schema: plan_schema.yaml
  • Runbooks: runbooks/

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.