<< All versions

Skill v1.0.0

currentLLM-judged scan90/100
haryy985/yunkun-console-preview/specializedarget-profiling
──Details
PublishedOctober 23, 2026 at 08:44 PM
Content Hashsha256:a4ac56d712c304da...
Git SHA49d280ffad42
──Files
Files (1 file, 1.7 KB)
SKILL.md1.7 KBactive
SKILL.md · 65 lines · 1.7 KB

version: "1.0.0" name: target-profiling description: > Target inventory, technology graph, trust-boundary, and attack-surface prioritization lane for authorized reviews.


Target Profiling Lane

This lane converts collection records into a bounded target profile. It does not replace collection and it does not claim a vulnerability from exposure alone.

Profile Schema

text
target_id:
scope:
domains:
hosts:
applications:
route_families:
identities:
tenants:
technologies:
data_classes:
trust_boundaries:
exclusions:
sources:
last_seen:
confidence:
open_questions:

Build Process

  1. Merge records by canonical host, route, application, and object key.
  2. Preserve every source, timestamp, query, and confidence value.
  3. Separate direct evidence from inference.
  4. Map public entry points to authentication, API, storage, job, and

administrative boundaries.

  1. Rank only in-scope assets by exposure, privilege, data sensitivity, and

business value.

  1. Produce two or three concrete hypotheses with a stated stop condition.

Priority Signals

Prioritize an asset when several independent signals agree:

  • a route is live and tied to a known application;
  • an identity or tenant boundary is observable;
  • a sensitive operation has a stable baseline;
  • a high-value object or state transition is reachable;
  • a technology fingerprint is confirmed by response or source evidence.

Treat shared hosting, banners, open ports, and stale DNS as context only until the application relationship is confirmed.

Output

Return a profile, a ranked shortlist, missing data, and the next bounded collection or validation action. Link each candidate to the evidence record that created it.

All versions