Miroprism
PassAudited by VirusTotal on May 11, 2026.
Overview
Type: OpenClaw Skill Name: miroprism Version: 1.4.0 MiroPRISM is a sophisticated orchestration protocol for multi-round AI reviews, designed to reduce consensus bias. The skill demonstrates high security awareness by implementing a mandatory sanitization phase (Phase 2 in SKILL.md) that strips code blocks, URLs, and structured data from reviewer outputs to prevent prompt injection between rounds. It uses standard shell-based concurrency management (PID-based .lock files) and local logging (metrics.tsv) within its designated workspace, with no evidence of data exfiltration, unauthorized network access, or malicious intent.
Findings (0)
Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.
Private design, code, or security-review details may be shown to several model subagents instead of just one.
The design intentionally broadcasts Round 1 findings to multiple reviewers in Round 2, creating an inter-agent communication flow. This is central to the protocol and disclosed, but users should know their review material and findings are shared among subagents.
R2 — Every reviewer responds to every finding: AGREE (cite evidence), DISAGREE (cite contradiction), or UNCERTAIN
Use MiroPRISM only for material you are comfortable sending through a multi-agent review, and redact secrets before invoking it.
Review outputs and cited project details can remain in local analysis files after the run.
The protocol stores intermediate reviewer outputs, digest logs, digests, Round 2 outputs, and final synthesis files. This persistence is disclosed and useful for the review workflow, but those files may retain sensitive project details.
analysis/miroprism/
runs/<slug>/
r1-outputs/
R1-digest-log.md
r1-digest.md
r2-outputs/
archive/<slug>/
YYYY-MM-DD-review-N.mdTreat the analysis/miroprism directory as sensitive, avoid committing it accidentally, and delete or protect it when reviewing confidential artifacts.
A weak or misleading Round 1 finding could still influence later reviewers and the final synthesis.
The artifact itself acknowledges that the shared Round 2 digest can influence downstream reviewers. The protocol includes mitigations such as sanitization, randomization, and evidence requirements, so this is a disclosed design limitation rather than a mismatch.
Reduces cascade sycophancy — does not eliminate it; the shared digest can still anchor R2 reviewers
For high-stakes reviews, use the optional digest review pause and independently inspect final high-confidence findings before acting on them.