<< All versions

Skill v1.0.0

currentLLM-judged scan90/100
haryy985/yunkun-console-preview/src-report-paradigm
──Details
PublishedOctober 23, 2026 at 08:55 PM
Content Hashsha256:b68d1b4a346af319...
Git SHA49d280ffad42
──Files
Files (1 file, 2.2 KB)
SKILL.md2.2 KBactive
SKILL.md · 78 lines · 2.2 KB

version: "1.0.0" name: src-report-paradigm description: > Evidence, severity, reproduction, and submission-quality lane for authorized security findings.


Report and Evidence Lane

This lane decides whether a candidate is ready to report. It does not turn weak signals into findings. Use it after a specialist has produced a concrete baseline and impact record.

Required Structure

  1. Summary: one sentence naming the affected function and security effect.
  2. Scope: exact host, route family, identity, and authorization context.
  3. Prerequisites: only the minimum setup needed to reproduce.
  4. Reproduction: complete requests and responses with secrets redacted.
  5. Control: the normal request or expected behavior.
  6. Impact: the changed object, identity, tenant, or business state.
  7. Cleanup: restored state, deleted test data, and final-state check.
  8. Fix direction: server-side control that enforces the invariant.

Three Gates

Gate A: Scope

The asset and action are authorized, in scope, and within the request budget. The identity and test data are identified.

Gate B: Proof

The report contains a control, one bounded mutation, a reproducible response delta, and an observable security impact. A route name, version, scanner finding, or copied token is not proof by itself.

Gate C: Reproducibility

Another reviewer can reproduce the result from the written sequence without guessing omitted headers, cookies, bodies, object IDs, or state transitions.

Decision

Use one of these states:

  • confirmed: all gates pass and impact is demonstrated;
  • incomplete: one proof item is missing;
  • negative: the control and mutation show no security impact;
  • blocked: authorization, state, rate limit, or safety prevents a valid

decision.

Do not assign severity before Gate B. Keep the rating proportional to the demonstrated impact, affected identities, data sensitivity, and required conditions.

Output Record

text
title:
asset:
affected_function:
identity:
precondition:
control_request:
mutated_request:
observed_delta:
security_impact:
cleanup:
evidence_refs:
status:
severity:
fix_direction:

See references/accepted-report-templates.md for the stable report shape.

All versions