prism-converge

prism-converge is a command for coding agents from Ronkupper/PRISM. It costs 33 tokens per session (560 once invoked), scanned A, original, MIT.

A command for bringing results from a PRISM research pass back into its central Master record.

In plain words
What is it for?
Use it to verify a dispatch ID, absorb vendor responses, reconcile findings, and record what changed or remains next.
Why use it?
It prevents truncated, mismatched, or ambiguous results from being silently added to the wrong research task.

Command

Part of the prism plugin — 1 skill, 7 commands shipped together

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add commands/ronkupper/prism/prism-converge
Clone the repo
git clone --depth 1 https://github.com/Ronkupper/PRISM

Or install prism, the plugin that ships this one along with the rest of its 1 skill, 7 commands.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for prism-converge

README.md
[![agentmods](https://agentmods.dev/badge/commands/ronkupper/prism/prism-converge.svg)](https://agentmods.dev/commands/ronkupper/prism/prism-converge)
Your own site
<a href="https://agentmods.dev/commands/ronkupper/prism/prism-converge"><img src="https://agentmods.dev/badge/commands/ronkupper/prism/prism-converge.svg" alt="Measured on agentmods" height="20"></a>
Per session 33 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 560 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00033 $0.00560
Opus 5 $0.00016 $0.00280
Sonnet 5 $0.00007 $0.00112
Haiku 4.5 $0.00003 $0.00056

Measured 5d ago against content hash 234ca7e2f4ec, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

prism-converge scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 5d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

plugins/prism/commands/prism-converge.md · 44 lines

What it actually says

Converge the returns from a dispatched pass back into the PRISM Master — the return leg of the dispatch round-trip (the dispatch-consumer role).

The returns to absorb (a pasted vendor output, a file path, or empty to use what is attached / in the project):

$ARGUMENTS

Then proceed:

  1. Load the framework if it is not already active: read PRISM_core.md (naming PRISM also auto-triggers the prism Skill, which loads the same core). Fetch reference/dispatch-conventions.md — the convergence-side bundle.
  2. Locate the Master and the Dispatch register. Identify which dispatch these returns answer; if it is ambiguous or no register entry matches, surface that and ask rather than guessing.
  3. Run the transport-integrity check (Step 0) before absorbing anything. Confirm the ━━━ END PRISM DISPATCHED PASTE — <Dispatch ID> ━━━ sentinel is present and that its Dispatch ID matches the register entry. Halt on a missing sentinel, a Dispatch-ID mismatch, or a truncated return — a mis-pasted or wrong-pass return must never be silently absorbed.
  4. Absorb and reconcile (Layer-1). Fold the returns into the Master's findings with provenance; run the convergence Monitors that bear — M8 (Stale Source), M9 (Convergence Type Drift), M10 (Rerun) — and surface any M7 (Source Conflict) rather than silently resolving it.
  5. Write canonical state. Append the convergence delta to the Master, mark the dispatch returned in the register, and rewrite What's next so the next session resumes from the updated plan.

If the returns look incomplete, contaminated, or off-pass, stop and surface it — do not absorb partial or suspect output into the Master.

This command is a Claude-Skill convenience for the convergence return-leg; it changes no framework mechanic. Saying "converge these returns" in plain language is equivalent, and is the portable form on every other vendor. It is the return sibling of /prism-whats-next (which stages the outgoing dispatch).

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 5d ago First seen · 44 lines · 33 tokens per session scan A 234ca7e2f4ec

Subscribe to this mod's changes

prism-converge is a command published in the GitHub repository Ronkupper/PRISM (5 stars, last pushed 2mo ago), licensed MIT. It adds 33 tokens to every session and 560 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.