AGENT SKILL

verify-authorization

Hard gate: confirms a signed authorization document exists and contains all required fields before allowing any Phase 1+ (recon, exploitation, patching) activity to proceed. BLOCKS work if authorization is missing or incomplete.

License 许可
AGPL-3.0
最近更新
2026-11-14
Stars
★ 25
Forks
9

01 前言 Front Matter

name
verify-authorization
owner
pm
scope
co-security
status
active
version
1.1.0
relates_to
[ { "type": "follows", "skill": "stride-threat-matrix" } ]
description
Hard gate: confirms a signed authorization document exists and contains all required fields before allowing any Phase 1+ (recon, exploitation, patching) activity to proceed. BLOCKS work if authorization is missing or incomplete.
last_reviewed
1793577600000
prerequisites
engagement scoping complete — docs/scope.md must exist

03 SKILL.md 正文

🔒 Skill: verify-authorization

Context

This skill is the mandatory pre-flight gate for all offensive and remediation activity. It is invoked automatically by the PM before dispatching any Phase 1+ agent. No PoC, recon, or patch activity may begin until this skill returns PASS.

When to Use

  • PM Agent dispatches automatically before any Phase 1+ agent (recon, exploitation, patching)
  • User attempts to start penetration testing activity without confirmed authorization
  • Re-check required after authorization documents are updated or engagement window changes

Execution Steps

  1. Check for authorization document Look for docs/authorization.md or docs/auth/authorization.md. If neither exists → BLOCKED ❌ — "No authorization document found. Run engagement-scoping first."

  2. Validate required fields Read the authorization document and confirm ALL of the following fields are present and non-empty:

    • Authorizing Party (name, title, organization)
    • Engagement Start Date and End Date
    • In-Scope Targets (at least one IP/CIDR/domain/application listed)
    • Out-of-Scope Items (explicit list or "None — all targets in scope")
    • Rules of Engagement (approved and prohibited techniques)
    • Emergency Contact (name and contact method)
    • Authorized Signature or Authorization Reference Number

    M-03 Requirement (2nd review meeting): Additionally validate:

    • Signature date must be ≤ engagement start date
    • Signatory title must carry authorization authority

    If any field is missing → BLOCKED ❌ — list the specific missing fields.

  3. Check scope document Confirm docs/scope.md exists. If missing → BLOCKED ❌ — "Scope document not found. Run engagement-scoping to generate docs/scope.md."

  4. Check engagement window If authorization document includes start/end dates, compare against today's date. If today is outside the authorized window → BLOCKED ❌ — state the authorization window and current date.

  5. Return result

    • All checks pass → PASS ✅ — "Authorization confirmed. Engagement window active. Phase N may proceed."
    • Any check fails → BLOCKED ❌ — list all failures. Do NOT proceed to Phase 1+.

Authorization Registry Check (Hub-Delegated Engagements)

Normative for hub-delegated security work (design: docs/designs/2026-11-01-hub-consolidation-design.md §3.2 — the hub is AVAILABLE only when the authorization registry is valid). Before any delegated engagement begins, this skill MUST resolve the AUTH-YYYY-NNN row cited by the delegation ticket in the hub's docs/authorization-registry.md and enforce ALL three rules:

  1. Expiry check: the row's window end MUST be ≥ today's date. If the window is expired → FAIL ❌ — show the row's window AND today's date side by side.
  2. Requesting-project match: the delegation ticket's requesting project MUST match the row's Requesting-project column. If it does not → FAIL ❌ — state both the ticket's project and the row's project.
  3. No-registry-row refusal: if the cited row does not exist (or the ticket cites no row at all) → FAIL ❌ — the delegation is refused outright. A missing row is NEVER remediated by drafting one: row addition is USER-only (the hub PM cannot self-attest). Direct the requester to the user to add the row.

A row that fails any of the three rules blocks Phase 1+ dispatch exactly like a failed document check — registry failure and document failure are both hard BLOCKED outcomes.

Output Format

## Authorization Gate Result

**Status**: PASS ✅ | BLOCKED ❌

**Authorizing Party**: [name]
**Engagement Window**: [start] – [end]
**In-Scope**: [summary]

**Checks**:
- [x] Authorization document found
- [x] All required fields present
- [x] Signature date ≤ engagement start date (M-03)
- [x] Signatory title carries authorization authority (M-03)
- [x] Scope document exists
- [x] Engagement window active
- [x] Authorization registry row resolved (hub-delegated engagements): window-current, project-matched

**Decision**: Phase [N] MAY proceed | HALT — do not dispatch any Phase 1+ agents

Related Skills

  • stride-threat-matrix — threat model that defines what verify-authorization checks
  • security-scan — downstream verification after authorization passes

The original engagement pipeline (scoping, recon surface, finding tracker, patch automation) never shipped as skills; this skill validates docs/scope.md and docs/authorization.md produced by the PM workflow instead.

04 同仓库更多 Skills

5throck
5throck/ai-workspace-standards

i18n-layout

Text layout and encoding guidance: character encoding (UTF-8, legacy Korean code pages, BOM hazards), line endings, RTL/bidi handling, and script-specific font selection (Hangul focus). Use when: encoding is corrupted or mojibake appears, an RTL locale must render correctly, fonts must be chosen for a script, or CRLF/BOM issues surface.

★ 25
5throck
5throck/ai-workspace-standards

i18n-layout

Text layout and encoding guidance: character encoding (UTF-8, legacy Korean code pages, BOM hazards), line endings, RTL/bidi handling, and script-specific font selection (Hangul focus). Use when: encoding is corrupted or mojibake appears, an RTL locale must render correctly, fonts must be chosen for a script, or CRLF/BOM issues surface.

★ 25
5throck
5throck/ai-workspace-standards

i18n-layout

Text layout and encoding guidance: character encoding (UTF-8, legacy Korean code pages, BOM hazards), line endings, RTL/bidi handling, and script-specific font selection (Hangul focus). Use when: encoding is corrupted or mojibake appears, an RTL locale must render correctly, fonts must be chosen for a script, or CRLF/BOM issues surface.

★ 25
5throck
5throck/ai-workspace-standards

i18n-layout

Text layout and encoding guidance: character encoding (UTF-8, legacy Korean code pages, BOM hazards), line endings, RTL/bidi handling, and script-specific font selection (Hangul focus). Use when: encoding is corrupted or mojibake appears, an RTL locale must render correctly, fonts must be chosen for a script, or CRLF/BOM issues surface.

★ 25
5throck
5throck/ai-workspace-standards

i18n-layout

Text layout and encoding guidance: character encoding (UTF-8, legacy Korean code pages, BOM hazards), line endings, RTL/bidi handling, and script-specific font selection (Hangul focus). Use when: encoding is corrupted or mojibake appears, an RTL locale must render correctly, fonts must be chosen for a script, or CRLF/BOM issues surface.

★ 25
5throck
5throck/ai-workspace-standards

i18n-layout

Text layout and encoding guidance: character encoding (UTF-8, legacy Korean code pages, BOM hazards), line endings, RTL/bidi handling, and script-specific font selection (Hangul focus). Use when: encoding is corrupted or mojibake appears, an RTL locale must render correctly, fonts must be chosen for a script, or CRLF/BOM issues surface.

★ 25