Skip to content

Extract vuln-assessment-methodology into standalone capability#40

Merged
mkultraWasHere merged 9 commits into
mainfrom
improvements/dotnet-rev
Jun 6, 2026
Merged

Extract vuln-assessment-methodology into standalone capability#40
mkultraWasHere merged 9 commits into
mainfrom
improvements/dotnet-rev

Conversation

@mkultraWasHere
Copy link
Copy Markdown
Contributor

Generic vulnerability assessment guidance was embedded in the dotnet-reversing skill. This extracts it into a reusable capability and wires it into web-security.

Added

  • vuln-assessment-methodology capability — severity matrix, disprove-first rules, anti-patterns table, reporting standards. No tools/agents/workers, just a single skill for any security capability to load.

Changed

  • dotnet-reversing skill now references vuln-assessment-methodology for generic methodology while retaining .NET-specific patterns (MarkupString, BinaryFormatter, JWT ReadToken, etc.), the severity table inline for standalone viability, and a compact pre-report checklist
  • mcr-analysis skill adds "Not All MCR Images Are .NET" fallback guidance, MCR-specific attack surface targets (ONNX, ANSI parsers, Protobuf, URL parsers), repo prioritization tiers, and subagent delegation instructions referencing both skills
  • web-security agent references the methodology skill in Evidence Standards
  • vuln-critic skill references the methodology's severity matrix and anti-patterns table

Notes

  • The dotnet-reversing skill duplicates the severity table from the methodology so it works standalone without the methodology loaded — intentional tradeoff

mkultraWasHere and others added 8 commits June 5, 2026 18:35
Generic vulnerability assessment guidance (source-to-sink tracing,
disprove-first analysis, severity matrix, quality checklist, anti-patterns)
was embedded in the dotnet-reversing skill. Extract it into a new
vuln-assessment-methodology capability so any security capability can
load it. dotnet-reversing and mcr-analysis now reference it while
retaining .NET-specific patterns and a compact inline severity table
for standalone viability.

Co-Authored-By: Claude <noreply@anthropic.com>
…severity example

- vuln-assessment-methodology: consolidate severity guidance (drop
  redundant prose lists, merge severity guide into rule 3, remove
  quality checklist that restated the rules, trim reporting standards)
- dotnet-reversing: fix report_finding example to use criticality
  "critical" for hardcoded credential, remove low-signal weak crypto
  pattern, fix stale "quality checklist" reference
- mcr-analysis: fix stale "quality checklist" reference, deduplicate
  dll_only fallback between sections

Co-Authored-By: Claude <noreply@anthropic.com>
…d vuln-critic

Add methodology skill reference to the web-security agent's Evidence
Standards section and to vuln-critic's opening paragraph. Gives the
pentesting pipeline access to the severity matrix, disprove-first
discipline, and anti-patterns table.

Co-Authored-By: Claude <noreply@anthropic.com>
…sis, and reporting rigor

Add hard rule 9 (attack chain analysis), three-tier confidence levels
(Confirmed/Probable/Suspected), opt-in PoC validation, and expand
reporting standards with CWE classification, actionable remediation,
root-cause deduplication, and scope documentation requirements.

Co-Authored-By: Claude <noreply@anthropic.com>
…n reporting

Rule 7: generalize to "security-sensitive sink" with examples instead of
exhaustive list. Rule 9: clarify chain is reported alongside individual
findings. Anti-pattern: align terminology with rule 9.

Co-Authored-By: Claude <noreply@anthropic.com>
…all finding agents

All 7 agents that report findings now load the methodology skill for
source-to-sink tracing, disprove-first analysis, confidence levels,
severity calibration, and reporting standards. attack-surface-mapper
is excluded as it does recon only.

Co-Authored-By: Claude <noreply@anthropic.com>
…logy in all finding agents"

This reverts commit 447d178.
…n and keywords

Description now reflects the full scope: confidence levels, chain
analysis, CWE mapping, remediation standards, dedup, scope, and
opt-in PoC validation.

Co-Authored-By: Claude <noreply@anthropic.com>
@mkultraWasHere mkultraWasHere marked this pull request as ready for review June 6, 2026 02:00
@mkultraWasHere mkultraWasHere requested a review from Copilot June 6, 2026 02:00
Copy link
Copy Markdown

Copilot AI left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Pull request overview

This PR extracts cross-cutting vulnerability assessment guidance into a new standalone vuln-assessment-methodology capability and wires it into existing security-focused skills/agents so multiple capabilities can share consistent severity calibration, disprove-first discipline, and reporting standards.

Changes:

  • Added new vuln-assessment-methodology capability (single skill) defining rules, severity matrix, confidence levels, reporting standards, and anti-patterns.
  • Updated web-security agent and vuln-critic skill to reference/loading guidance for the methodology skill.
  • Updated .NET reversing and MCR analysis skills to reference the methodology and expanded guidance (including additional patterns and workflow notes).

Reviewed changes

Copilot reviewed 6 out of 6 changed files in this pull request and generated 1 comment.

Show a summary per file
File Description
capabilities/web-security/skills/vuln-critic/SKILL.md References the new methodology skill as the source for severity/anti-patterns guidance.
capabilities/web-security/agents/web-security.md Adds instruction to load the methodology skill under Evidence Standards.
capabilities/vuln-assessment-methodology/skills/vuln-assessment-methodology/SKILL.md New methodology skill with hard rules, severity/confidence guidance, reporting standards, and anti-patterns.
capabilities/vuln-assessment-methodology/capability.yaml New capability definition/metadata for the methodology capability.
capabilities/dotnet-reversing/skills/mcr-analysis/SKILL.md Adds methodology-skill load guidance and expands MCR-specific analysis and delegation guidance.
capabilities/dotnet-reversing/skills/dotnet-reversing/SKILL.md Adds methodology references and expands vulnerability-pattern examples and reporting guidance.

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

Comment thread capabilities/dotnet-reversing/skills/dotnet-reversing/SKILL.md Outdated
…te methodology frontmatter

Path traversal "SAFE" example was bypassable via directory prefix
collision (/app vs /app2). Normalize baseDir with trailing separator
before StartsWith check. Also update vuln-assessment-methodology
skill frontmatter description to reflect expanded scope.

Co-Authored-By: Claude <noreply@anthropic.com>
@mkultraWasHere mkultraWasHere merged commit 94c671d into main Jun 6, 2026
3 of 4 checks passed
mkultraWasHere added a commit that referenced this pull request Jun 6, 2026
… (#41)

PR #40 updated skills and agents in both capabilities without bumping
their versions. dotnet-reversing 1.0.0 → 1.0.1, web-security 1.1.1 → 1.1.2.

Co-authored-by: Claude <noreply@anthropic.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants