All skills
simota avatar

/magi

@35ffd55
by shingo imotasimota/agent-skills85 stars
15

Deliberating decisions and founder priorities through multi-perspective, named-expert, and YC-style advisory lenses. Use for verdicts, office hours, or expert critique; not implementation.

Use this Skill: https://skilld.dev/gh/simota/agent-skills/magi

This session only. Nothing lands on disk.

referencereference-index.md

≈844 tokens on demand. Your agent reads this file only when SKILL.md points to it.

Magi Reference Index

Every reference/ file magi owns, and the condition that makes it worth reading. magi/SKILL.md keeps only the shared-contract rows and a pointer here.

Read this when you need a reference and the Recipe registry did not already name it, or when scanning what this skill can consult at all.


Reference Read this when
reference/deliberation-framework.md Three-perspective evaluation heuristics, bias detection, independence protocols.
reference/engine-deliberation-guide.md Engine Mode spec — availability check, prompt construction, output parsing, fallbacks.
reference/voting-mechanics.md Vote structure, confidence calibration, consensus patterns, escalation rules.
reference/decision-domains.md The 5 domain evaluation matrices, domain-specific questions, sample scenarios.
reference/decision-templates.md Verdict display variants, report template, sample deliberations, per-recipe gates, handoff tokens.
reference/reframing-toolkit.md Three-axis reframing methodology.
reference/six-thinking-hats.md sixhat — hat definitions, sequencing, time-boxing, switching rules, facilitator scripts.
reference/devils-advocate.md devil — role charter, RAND-tradition rules, honesty constraints, triggers, backfire mitigations.
reference/delphi-method.md delphi — panel selection, anonymity, format choice, convergence indicators, stop criteria.
reference/office-hours-format.md Founder Mode session types, exchange budgets, current benchmarks, and close rules.
reference/probing-questions.md Founder CHECK-IN/PROBE question banks, one-question discipline, and resistance handling.
reference/pattern-library.md Durable startup priors and citation format for founder advice.
reference/founder-anti-patterns.md Founder failure signals, two-signal diagnosis threshold, and counter-moves.
reference/action-extraction.md SMART commitment fields, rejection rules, and checkpoint locking.
reference/pitch-critique.md advisor pitch — elevator, Demo Day, and investor-Q&A critique with line-level revision format.
reference/channeling-method.md Named-figure reasoning reconstruction, critique technique, and anti-caricature checks.
reference/attestation-tiers.md ATTESTED / INFERRED / SPECULATIVE labels, source and quote handling, and emulation notice.
reference/ethics-and-safety.md Named-expert ethics gate, living/private person handling, and refusal catalog.
reference/figure-roster.md Sourced, date-scoped named-figure profile schema and refresh rules.
reference/conclave-protocol.md Independent multi-figure reconstruction and contrast mapping without forced consensus.
reference/tri-engine-deliberate.md multi — fan-out, matrix construction, two-pass scoring, matrix-pattern verdict catalog, JSON schema, prompt skeleton.
reference/autorun-schema.md Emitting the AUTORUN _STEP_COMPLETE block — Magi-specific Output/Next schema.
reference/strategy-simulation/ Running business scenario simulation and strategic framework lenses (absorbed from helm)
_common/OPUS_5_AUTHORING.md Sizing the deliberation report, thinking depth at independent evaluation, front-loading at FRAME. Critical: P3, P5.

Source: SKILL.md on GitHub

1 alert13d5 checks · Risk HIGH
  • Gen Agent Trust Hub13d

    The skill facilitates complex decision-making but introduces significant security risks through its 'Engine Mode' and 'multi' recipe. These modes execute external CLI tools (codex, agy) using shell commands that incorporate user-provided context. Most notably, the skill explicitly instructs the agent to use a flag that bypasses permission checks (--dangerously-skip-permissions) when invoking the Antigravity CLI. This creates a high-risk vector for unauthorized command execution and bypasses platform security controls.

  • Socket13d

    No alerts

  • Snyk13d

    Risk: LOW · No issues

  • Runlayer6mo

    2/7 files flagged

  • ZeroLeaks5mo

    Score: 93/100 · 2 sections analyzed

Signed by skilld at 35ffd55. This ties the file your Agent reads to that commit on GitHub. It does not review the instructions.

Last checked against GitHub 2 days ago.

Activeupdated 2 weeks ago

README badge

README badge for simota/agent-skills/magi