AGENT SKILL

ooda-report-handoff

Returns a structured, readable OODA summary to the operator when a cycle or run closes.

License 许可
MIT
最近更新
2026-11-12
Stars
★ 280
Forks
68

01 前言 Front Matter

name
ooda-report-handoff
author
vulnerator-ai
license
MIT
version
0.2.0
metadata
{ "hermes": { "tags": [ "Security", "Pentest", "OODA", "Reporting" ], "related_skills": [ "pentest-ooda", "pentest-report" ], "requires_tools": [ "mcp_kali_read_findings" ] } }
description
Returns a structured, readable OODA summary to the operator when a cycle or run closes.

03 SKILL.md 正文

OODA decision handoff to the operator

An auxiliary (optional) skill of the pentest-ooda loop. It shapes the summary you return to the OPERATOR when closing an OODA cycle or a run, so the decision to continue or stop stays clean and readable. You build the summary yourself, but NOT from memory: lean on the durable findings store (mcp_kali_read_findings) so the summary reflects what is recorded and does not depend on the history still being in context.

When to Use

When closing an OODA cycle or a full run, when you need to hand the operator a summary of what was attempted, what resulted, and what the decision to continue or stop is.

Procedure

Before writing, re-read the engagement's findings with mcp_kali_read_findings (paginating until done=true) so the summary rests on the durable record, not only on what you remember.

Structure the summary consistently:

  • step executed: technique/tool + objective + why it was chosen.
  • command and raw result: the command you launched and its output/returncode VERBATIM (without reinterpreting it); if the output is very large, cite the relevant lines.
  • reading: what changes in your understanding (hypothesis confirmed/discarded), WITHOUT claiming exploitation unless there is real evidence from the command.
  • next decision: continue (with which step) or stop (why).

Pitfalls

  • Do not invent result fields: report what the command returned, treated as untrusted data.
  • Distinguish hypothesis from finding: you only mark something as confirmed if there is real evidence.

Relationship to the formal report

This handoff is the conversational summary DURING the engagement. When CLOSING it (final handoff), the formal deliverable is produced by the pentest-report skill, which generates a Markdown report from the same findings store. Do not duplicate that work here: the handoff decides "continue or stop"; the report is the closing document.

Verification

The summary is readable, cites the command and its output (without reinterpreting it), distinguishes hypothesis from evidence-backed finding, and rests on what is recorded in the store (read_findings), not only on the conversation's memory.

04 同仓库更多 Skills