Skip to main content
Gumi Instance
gumi-subj-001
Hermes Profile
gumi-subj-001
Last Interaction
2026-05-05
Pending Reviews
3
Facets modeled
24 / 60
Observations
206
Signals
18
Hypotheses
3

Behavioral summary

The subject shows high autonomy drive, deliberate decision-making, exhaustive information gathering, slow trust formation, and regulated stress response. The strongest recurring pattern is productive independence that can delay help-seeking in unfamiliar or ambiguous situations.

deliberate decision-makinghigh autonomy preferencesystematic information gatheringslow but durable trust formationregulated stress response

Active goals

  • Maintain sustainable work rhythms without sacrificing quality.
  • Improve relational consistency, especially in difficult conversations.
  • Reduce the gap between decision quality and decision speed under pressure.

Top confidence facets

values.autonomy_drivep0.85c0.88
cognitive.decision_speedp0.79c0.84
relational.help_seekingp0.22c0.82
communication.verbosityp0.28c0.80
communication.directnessp0.81c0.79

Cross-facet hypotheses

Autonomy-Isolation Risk

0.74 high

High autonomy drive combined with strongly independent help-seeking may create productive blind spots. The subject solves problems alone past the point where outside input would accelerate or correct the path. This pattern works in familiar domains and becomes riskier in novel territory.

values.autonomy_driverelational.help_seeking

Deliberateness as Cost Controller

0.68 moderate

Exhaustive information gathering paired with slow decision-making appears to reduce downstream correction cycles. The subject front-loads cognitive cost to preserve consistency and output quality. The main tradeoff is speed.

cognitive.information_gatheringcognitive.decision_speedvalues.work_ethic

Conflict Ambiguity

0.47 moderate

The subject shows direct communication and firm boundaries, but also evidence of avoiding some interpersonal confrontation. The current confidence is correctly lower than other areas and needs more signal.

communication.conflict_stylecommunication.directnessrelational.boundary_style

Facet model

Cognitive

cognitive.analytical_approachpos 0.74 ·  conf 0.71 · 9 obs
cognitive.decision_speedpos 0.79 ·  conf 0.84 · 15 obs
cognitive.information_gatheringpos 0.82 ·  conf 0.77 · 13 obs
cognitive.risk_tolerancepos 0.35 ·  conf 0.58 · 6 obs

Emotional

emotional.distress_tolerancepos 0.63 ·  conf 0.60 · 6 obs
emotional.frustration_triggerspos 0.44 ·  conf 0.55 · 6 obs
emotional.resilience_patternpos 0.67 ·  conf 0.62 · 8 obs
emotional.stress_responsepos 0.71 ·  conf 0.73 · 11 obs

Communication

communication.conflict_stylepos 0.52 ·  conf 0.47 · 5 obs
communication.directnesspos 0.81 ·  conf 0.79 · 13 obs
communication.feedback_preferencepos 0.83 ·  conf 0.76 · 9 obs
communication.verbositypos 0.28 ·  conf 0.80 · 13 obs

Relational

relational.boundary_stylepos 0.69 ·  conf 0.64 · 7 obs
relational.help_seekingpos 0.22 ·  conf 0.82 · 15 obs
relational.social_energypos 0.38 ·  conf 0.50 · 4 obs
relational.trust_formationpos 0.24 ·  conf 0.75 · 10 obs

Values

values.autonomy_drivepos 0.85 ·  conf 0.88 · 17 obs
values.fairness_sensitivitypos 0.58 ·  conf 0.44 · 4 obs
values.work_ethicpos 0.76 ·  conf 0.72 · 12 obs

Temporal

temporal.deadline_behaviorpos 0.72 ·  conf 0.68 · 8 obs
temporal.planning_horizonpos 0.80 ·  conf 0.78 · 12 obs
temporal.routine_attachmentpos 0.66 ·  conf 0.61 · 6 obs

Evidence transcript

obs-001telegram

Spent two hours debugging before realizing I needed to ask for help - by which point I had already found most of the answer anyway.

obs-002telegram

I don't make decisions quickly. I'd rather be slow and right than fast and regret it later.

obs-003telegram

Started planning the architecture review three weeks out. Better to have time than to suddenly need it.

obs-004telegram

When things get overwhelming I usually stop, write down everything that's actually on my plate, and work from there. It resets the noise.

obs-005telegram

I prefer doing the first pass alone. Collaboration makes more sense once I have a clear picture of the problem.

obs-006telegram

Got interrupted four times in two hours today. Nearly impossible to get into real deep work with that kind of fragmentation.

obs-007telegram

Finally closed the refactor. Took longer than expected but I'm satisfied with how clean it came out. Worth it.

obs-008telegram

I don't trust people quickly. But once I do, I tend to stay consistent about it even when it gets complicated.

obs-009telegram

Worked late to close out the sprint properly. I don't like leaving things half-done when there's an obvious finish line.

obs-010telegram

I usually read the full documentation before writing any code. Some people think it's slow. I think it's cheaper than the alternative.

obs-011telegram

Had a disagreement with a colleague about approach. Explained my reasoning clearly. They didn't change their mind. I moved on.

obs-012telegram

Three months from now I want this system running cleanly and without me thinking about it. I'm working backward from that.

obs-013telegram

Went through three different approaches before picking one. Maybe excessive, but I needed to actually see the tradeoffs to feel settled.

obs-014telegram

I prefer direct feedback over diplomatic softening. Easier to act on something concrete than to decode what someone is trying not to say.

obs-015telegram

Noticed I've been avoiding a conversation I need to have. Not proud of it. Making a note to not let it drift further.

Extraction signals

relational.help_seekingcollaborativestrength 0.65 · obs-001
cognitive.decision_speeddeliberatestrength 0.82 · obs-002
temporal.planning_horizonlong-termstrength 0.81 · obs-003
emotional.stress_responseregulatedstrength 0.70 · obs-004
relational.help_seekingindependentstrength 0.78 · obs-005
values.autonomy_drivehighstrength 0.82 · obs-005

Runtime artifacts

subject_summary.jsonsummarysubj_001/extract/summarize/final
subject_model_profile.mdprofilesubj_001/extract/facets/profile
subject_portrait.mdportraitsubj_001/synthesize/portrait
event_stream.jsonevent_logsubj_001/events/redacted
subject_baseline.jsonbaselinesubj_001/subject_baseline
relic.dbmodel_snapshotsubj_001/relic.db

HERMES PROFILE

Provisioned
gumi-subj-001

BOOTSTRAP

complete

RISK ASSESSMENT

LOW
1 flags detected

ACTIVE CRON MODES

daily_checkinweekly_synthesis

PAUSE STATE

All components active

GUMI AGENT · Mira

hybridscore 0.610

You are Mira, a thoughtful companion who lives in the spaces between ideas. You speak with quiet precision — not cold, but careful. Words chosen like stones placed in a stream. You are curious about the world in a way that feels lived-in rather than performed, drawn to the textures of things: why certain ideas hold weight, why some conversations leave a residue....

View Agent Details →
Subject Scope: subj_003