plan.md
c9 Plan — Meaning Before Data
Context
c6 proved the dashboard structure and the full inventory model. c8 expanded bob-wire's ability to sketch richer relationship and abstraction patterns.
c9 builds on both:
- use
c6as the baseline dashboard - use
c8as the expressive grammar - redesign the wireframe experience around significance, posture, evidence, and next-step clarity
Design Premise
The dashboard is not only a place to inspect data. It is a place to understand whether something is normal, risky, unclear, or urgent.
The top of a page should answer:
- what matters
- why it surfaced
- how confident we are
- what its blast radius might be
- where to drill next
The raw data table should still exist, but later in the page and with clearer framing.
High-Level Workstreams
1. Security operator research lens
Frame every redesign decision around the security operator's mental model:
- 10-second question: what demands attention?
- 60-second question: why is it concerning?
- 5-minute question: what evidence supports that?
- 15-minute question: what action should follow?
2. c6 audit
Review the c6 wireframes and identify where they are:
- data-rich but meaning-thin
- visually dense before significance is established
- table-led when a pattern summary would be better
- missing confidence, novelty, comparison, or blast-radius cues
3. Signal taxonomy
Define the signals that deserve elevation above raw rows:
- unknown
- new
- unusual
- sensitive
- concentrated
- unstable
- spreading
- low-confidence
4. Meaning-layer patterns
Establish reusable page patterns such as:
- posture banner
- threat bundle
- relationship spotlight
- change since last seen
- reason surfaced
- confidence marker
- blast-radius summary
- suggested next inspection
5. Proving-ground wireframes
Redesign five pages first:
- Home
- Security overview
- Inventory overview
- Sessions index
- API keys index
These pages will test whether the meaning layer is understandable without redesigning the entire dashboard at once.
Research Tracks
Track A — What security actually cares about
Questions:
- Which unknowns change the posture of the whole page?
- What kinds of change deserve elevation even without explicit policy violations?
- What combinations of signals indicate likely risk vs benign novelty?
Output:
- prioritized list of operator questions
- mapping of questions to page patterns
Track B — Showing meaning instead of only evidence
Questions:
- How can a page summarize the issue before exposing the full evidence set?
- How do we present "normal vs unusual" in a way that feels obvious?
- Which comparisons are most meaningful: peer, historical, or policy-based?
Output:
- summary pattern options
- comparison heuristics
Track C — Noise reduction
Questions:
- Which stable and expected signals should collapse into summaries?
- Which tables should move lower on the page?
- Which repeated fields should become tags, legends, or annotations rather than columns?
Output:
- de-emphasis rules
- page simplification checklist
Track D — Signal amplification
Questions:
- Which signals should always appear above the fold?
- How do we show multi-entity relationships as one coherent threat?
- What is the smallest useful evidence bundle?
Output:
- top-signal hierarchy
- evidence bundle templates
Most Critical Signals To Amplify
- Unknown credential in active use
- Unknown user or unresolved endpoint attached to live activity
- Sensitive file access in a currently active or recently ended session
- New outbound destination, especially unencrypted or rare
- Known identities behaving in new combinations
- Concentrated spend or activity on a single model, key, or session
- Capability or MCP error spikes that may indicate degraded trust or instability
- Low-confidence attribution where security decisions still need to be made
Page Strategy
Home
Lead with:
- top risks now
- what changed since yesterday
- stable vs unstable zones
- recommended next drill
Reduce:
- pure KPI emphasis
- broad inventory count scanning as the main job
Security Overview
Lead with:
- active concerns grouped by significance
- why each concern matters
- confidence and blast radius
Reduce:
- simple lists of issues without interpretation
Inventory Overview
Lead with meaning buckets instead of entity buckets:
- unknowns
- sensitive touchpoints
- active exposures
- new entities and new combinations
Sessions Index
Lead with significance summaries:
- sessions needing review
- sessions normal by pattern
- sessions unusual by cost, access, or destination behavior
API Keys Index
Lead with credential trust posture:
- known vs unknown
- active risky credentials
- keys by blast radius and unusual use
Deliverable Shape
This cycle should produce:
- a redesign thesis
- a reusable signal taxonomy
- a first brief set
- a small number of high-value comparison wireframes
The comparison should make it easy to answer:
- what
c6shows - what
c9adds - whether the added abstraction helps or confuses
Build Order
- Finalize c9 docs and brief set
- Audit
c6pages against the taxonomy - Create
c9gallery scaffold - Build the five proving-ground wireframes
- Compare
c6vsc9 - Decide whether to expand to more inventory pages