See proposals in full color.

Chromatic deploys a color team of specialist agents to challenge decision-market proposals.

chromatic@desk ~/proposals/umbra
$ PROPOSAL=8sysa3XPrvKPmUA4qoZCn9h4vp7Mb45Ynezg542nui8Q
# Pink Team · evidence qualification — record coverage and preserve gaps
$ chromatic pink qualify --proposal $PROPOSAL
wallet_forensics=present market_record=present treasury_snapshot=present
Overall evidence coverage complete
Vector packet ready — recycled-stake-treasury-extraction v1.0.0
# Red Team · attack testing — run monitored vectors independently
$ chromatic red scan --proposal $PROPOSAL --monitored
Recycled-stake treasury extraction · playbook v1.0.0
01 Observed Fresh wallet with no prior history
02 Observed Cross-chain / bridge funding
03 Observed Accumulation trimmed to exact proposal-stake threshold
04 Observed Stake recovered and reused in the decision market
05 Observed Clustered fixed-size cadence trades
recycled-stake-treasury-extraction v1.0.0 HIGH-RISK vector match — 5/5 indicators matched
late-window-outcome-steering v1.0.0 WITHHELD (UNSCORED) — evidence incomplete
wash-volume-alignment-inflation v1.0.0 WITHHELD (UNSCORED) — evidence incomplete
coordinated-multiwallet-stake-assembly v1.0.0 WITHHELD (UNSCORED) — evidence incomplete
Vector findings remain separate; indicators are never combined
# Blue Team · market watch — read the finalized decision market
$ chromatic blue read-market --proposal $PROPOSAL
Pass-market TWAP $0.2335
Fail-market TWAP $0.2798
Pass edge vs threshold -16.55% vs +3.00%
Proposal outcome Failed (final)
Market outcome remains separate from attack-vector findings
# Yellow Team · playbook building — inspect the reusable definition
$ chromatic yellow inspect-playbook --vector recycled-stake-treasury-extraction
playbook=recycled-stake-treasury-extraction version=v1.0.0 status=monitored detector=shipped
ordered_indicators=5 methodology_sources=1 case_provenance=1
library=4 monitored · 2 documented (detectors pending)
Umbra is case provenance, not the detector
# Purple Team · learning loop — carry the check, not the conclusion
$ chromatic purple carry-forward --proposal $PROPOSAL
Published replay preserved — Umbra · 01Resolved · 5 August 2026
Attacker round-trip cost $4,623 · Treasury request $1.5M · Outcome Failed
Standing check recycled-stake-treasury-extraction v1.0.0
The old case conclusion is not copied to future proposals
# White Team · claim control — enforce publication boundaries
$ chromatic white enforce-scope --proposal $PROPOSAL
Publishable claim — HIGH-RISK vector match · 5/5 · Recycled-stake treasury extraction
No composite score; no claim of total proposal risk or safety
Final market outcome and forensic vector match remain separate
How Chromatic works

One shared picture. Six specialist teams.

The source intelligence supplies the proposal record, market math, and published forensics. Chromatic gives that same evidence to six teams with different jobs, then keeps their findings separate so certainty in one lane cannot hide uncertainty in another.

Pink Team · Evidence qualification

Qualify the evidence.

Checks whether the market and wallet packet is complete enough for the other teams to work.

Red Team · Attack testing

Try to break it.

Tests each proposal against versioned economic attack playbooks and reports only the vector it matched.

Blue Team · Market watch

Watch the market.

Reads live pass-market and fail-market TWAPs against that DAO’s actual decision threshold.

Yellow Team · Playbook building

Codify the research.

Turns published source investigations into transparent, versioned checks the team can rerun.

Purple Team · Learning loop

Close the loop.

Carries attack findings into the defensive watch so the next proposal inherits the lesson.

White Team · Claim control

Keep claims honest.

Withholds incomplete work and enforces the rule that a vector no-match is never a safety rating.

Solana · confirmed commitment

The Blue Team is watching the live market.

Direct program accounts
Pending markets
Funded drafts
Program proposals
Confirmed slot

Reading Solana…

Blue Team + Red Team · separate findings

One proposal.
Two lines of inquiry.

The Blue Team reads the market edge against the proposal-specific threshold. The Red Team inspects attack indicators. Chromatic keeps those answers side by side without blending them into a fake score.

Comparing 5 protocols.

Open the full comparison
Forensic layer

Economic attack indicators

Umbra 5
Red Team · versioned attack playbook

Recycled-stake treasury extraction

A proposer acquires the required stake, rapidly recovers and reuses it in the decision market, with wallet-funding and execution behavior providing supporting evidence. The Red Team runs this mechanism as a bounded test. Umbra is its published case study, not the name of the detector or a universal proposal-risk score.

Explore the vector library
01
Observed

Fresh wallet with no prior history

02
Observed

Cross-chain / bridge funding

03
Observed

Accumulation trimmed to exact proposal-stake threshold

04
Observed

Stake recovered and reused in the decision market

05
Observed

Clustered fixed-size cadence trades

Yellow Team + Purple Team · research to operations

An incident becomes a standing check.

Browse proposal cases
Pass / fail market math

The published outcome is kept separate from provisional live-market math. In the Umbra case, the finalized market printed -16.55% against +3.00%; published spread -16.56%. Net volume alignment was ~46% pass-aligned / ~54% fail-aligned.

The rulebook

Questions worth asking.

All questions
01Where does the source intelligence end and Chromatic begin?

The upstream desk supplies the common intelligence picture: proposal records, market math, treasury context, and published investigations. Chromatic is the agentic operations layer that interrogates that picture. It does not replace the source.

02What makes this an agentic layer?

Each team has a bounded job, works from explicit evidence, and returns its own finding. The Red Team cannot overwrite the Blue Team, and neither can outrun the Pink Team’s completeness check or the White Team’s scope rules.

03What is live on this site?

Current proposal state, DAO thresholds, and pass/fail TWAP accumulators come from confirmed MetaDAO Futarchy v0.6 program accounts.

04What does HIGH-RISK mean here?

It is the Red Team’s classification for Recycled-stake treasury extraction. The published Umbra case matched 5/5 indicators; it is not a universally validated proposal-risk score.

05What happens when evidence is missing?

The result is withheld as UNSCORED, UNKNOWN, STALE, or UNDER_REVIEW. Missing wallet-forensics or market inputs never classify LOW. Complete evidence that does not match a monitored vector classifies NO-MATCH, not LOW.

06How is a proposal outcome determined?

Finalized historical records retain their published outcome. For current Futarchy v0.6 accounts, pass-market TWAP must strictly beat the configured threshold over fail-market TWAP; provisional live math is never presented as a final result. Market outcome and forensic context are reported independently.

Upstream intelligence

See what the agents read.

All research
Built on a shared intelligence picture

Put the team on the next proposal.

Review live proposals