Skip to main content

Documentation Index

Fetch the complete documentation index at: https://mintlify.com/XxYouDeaDPunKxX/PA-PVP-mini/llms.txt

Use this file to discover all available pages before exploring further.

You do not need a prior PA-PVP Mini round to use PA-PVP Mini in Round 2+ mode. If you already have a structured review, a findings list from a colleague, linter output, a security audit, or any other prior analysis, you can feed it directly into the protocol as PREVIOUS_ROUND_OUTPUT. PA-PVP Mini will treat it as audit material — checking it against the original artifact rather than simply accepting its conclusions.

​
What qualifies as valid prior material

PREVIOUS_ROUND_OUTPUT accepts any structured prior analysis, including:

Previous PA-PVP rounds

Output from an earlier PA-PVP Mini session, whether run by you or another person.

Findings lists from colleagues

Peer review notes, code review comments, or structured feedback from a team member.

Tool output

Linter results, static analysis output, security audit findings, or any structured diagnostic output.

Notes and informal reviews

Structured notes, decision memos with objections listed, or findings from a prior meeting.
The material does not need to follow PA-PVP Mini’s output format. What matters is that it contains reviewable claims about the artifact that the protocol can check.

​
How PA-PVP Mini treats prior material

ORIGINAL_ARTIFACT is always the primary source of truth. PREVIOUS_ROUND_OUTPUT — regardless of where it came from — is audit material. PA-PVP Mini runs a full independent review of the artifact first, then cross-checks the prior findings against that review. It will confirm what holds up, challenge what doesn’t, and surface anything the prior review missed. This means the protocol never inherits the prior review’s conclusions uncritically. A finding from a colleague or a tool is treated the same way as a finding from a previous PA-PVP round: it must be defensible against the artifact or it gets challenged.

​
Input format

Use the same two-block format regardless of where your prior material came from: 
ORIGINAL_ARTIFACT:
[your material]

PREVIOUS_ROUND_OUTPUT:
[your existing review or findings]
Paste the full contents of pa_pvp mini v2.3.txt into the chat first to activate the protocol, then send this input.

​
What the AI does with it

PA-PVP Mini executes Round 2+ behavior in strict order:
  1. Produces a complete independent review of ORIGINAL_ARTIFACT — without referencing the prior material at all
  2. Cross-checks the prior findings against that independent review
  3. Returns a merged output with CONFIRMED, CHALLENGED, and NEW labels on every finding, plus a MERGE STATUS block
The result is a structured audit of both the artifact and the prior review — making explicit what the prior review got right, what it missed, and where it overreached.
Supplying PREVIOUS_ROUND_OUTPUT automatically triggers Round 2+ behavior. You do not need to set a round number manually. If your prior material contains explicit round numbers, PA-PVP Mini will set the current round to the highest prior round number plus one. If no round numbers are present, it defaults to Round 2.
If you have only a short piece of carry-over context — a constraint that was added after the review, a known exception, or a brief clarification — use MEMORY_NOTE instead of PREVIOUS_ROUND_OUTPUT. It is treated as audit material in the same way, but is the right block for context that is not itself a structured review.
ORIGINAL_ARTIFACT:
[your material]

MEMORY_NOTE:
[short carry-over context]

Build docs developers (and LLMs) love

Get started for freeTalk to us