idea-auditor

Plugin idea-auditor from marketplace nsalvacao-claude-code-plugins in nsalvacao/nsalvacao-claude-code-plugins. Commands read from the repo's manifest — 1 month ago.

Skills (6)

• idea-auditor-calibrate

  Guides a deliberate, evidence-grounded recalibration of one or more dimension scores. Produces a calibration log (calibration-log-YYYYMMDD.md) with before/after scores, the evidence that changed, and the rationale. Enforces anti-p-hacking rules — score adjustments require new evidence, not score-sho

• idea-auditor-drill

  Deep-dives into a single scoring dimension using the appropriate specialist agent. Produces a focused analysis with score_bruto, evidence gaps, and ≥3 targeted experiments with stop rules. Invoke when a specific dimension is weak in the scorecard or when the user asks to analyze wedge, friction, loo

• idea-auditor-migrate

  Assesses switching costs and migration feasibility for infrastructure forks, standards adoptions, and migration-heavy products. Decomposes switching costs into technical, learning, transaction, and lock-in components using the Klemperer taxonomy. Produces a migration scorecard with diff surface meas

• idea-auditor-report

  Generates a full consolidated report from an existing scorecard and evidence JSON. Covers all dimensions, final decision, top 3 blockers with improvement paths, and next tests. Invoke when the user asks for a full report, executive summary, or runs /idea-auditor:report.

• idea-auditor-score

  Runs the idea-auditor scoring pipeline on an IDEA.md or IDEA.json file. Validates inputs, grades evidence from STATE/, calculates confidence-weighted dimension scores using scientific frameworks (JTBD, TAM, Rogers, Mayer), and produces a scorecard with a PROCEED/ITERATE/KILL decision, top blockers,

• idea-auditor-watch

  Explains the passive watch mode that automatically snapshots high-signal files (IDEA.*, BLUEPRINT.*, README.md, SECURITY.md, CHANGELOG.md) after every write. Shows which files are being observed, what is stored in STATE/.snapshots/, and how to use diff_scorecards.py to detect regressions between sna

Slash commands (7)

• /idea-auditor:calibrate

  Deliberately recalibrates one or more dimension scores after new evidence arrives. Enforces anti-p-hacking rules — every score adjustment must cite a specific new evidence item dated after the last scoring run. Produces a calibration log (REPORTS/calibration-log-YYYYMMDD.md) with before/after scores

• /idea-auditor:drill

  Deep-dive into a single scoring dimension using its specialist agent. Invokes the appropriate dimension agent (wedge-researcher, friction-analyst, loop-designer, timing-scout, or trust-auditor), grades available evidence, and produces a focused report with score_bruto, evidence gaps, and ≥3 experime

• /idea-auditor:migrate

  Assesses switching costs and migration feasibility for infrastructure forks and standards-adoption products (Infra_Fork_Standard mode). Decomposes switching costs using Klemperer taxonomy — technical, learning, transaction, lock-in — and measures diff surface as the primary quantitative proxy. Produ

• /idea-auditor:report

  Generates a full consolidated Markdown report from the latest scorecard and evidence JSON. Covers all dimensions, final decision, top 3 blockers with improvement paths, and next steps. Requires a scorecard to exist in REPORTS/ (run /idea-auditor:score first).

• /idea-auditor:score

  Run the idea-auditor scoring pipeline on an IDEA.md or IDEA.json file. Validates inputs, grades evidence from STATE/, and produces a PROCEED/ITERATE/KILL scorecard with top blockers and next tests. Usage: /idea-auditor:score <path> [--mode OSS_CLI|B2B_SaaS|Consumer_Viral|Infra_Fork_Standard]

• /idea-auditor:tests

  Generate a structured experiment plan for an idea or MVP with stop rules, proxy metrics, and minimum sample sizes. Targets the weakest or most uncertain dimensions from a prior scorecard. Usage: /idea-auditor:tests <path>

• /idea-auditor:watch

  Shows the watch mode status — which files are being observed passively, what is stored in STATE/.snapshots/, and how to use diff_scorecards.py to compare scorecards and detect regressions. Watch mode is always active while the plugin is installed; it snapshots high-signal files (IDEA.*, BLUEPRINT.*,

Subagents (10)

• idea-auditor-competitor-mapper

  Use this agent when competitive landscape analysis is needed — mapping alternatives, normalizing them to the same proxy metrics, and identifying competitive heat in the idea's target segment. Does not produce a score_bruto for the scorecard; instead produces a normalized competitor matrix that infor

• idea-auditor-evidence-grader

  Use this agent when evidence needs to be graded before scoring — classifies each item into 4 confidence components (source diversity, recency, commitment, consistency) and computes ConfDim. Never invents data; missing evidence results in score=null, not a guess. <example> Context: User wants to grad

• idea-auditor-experiment-designer

  Use this agent when an experiment plan needs to be designed with rigorous stop rules, minimum sample sizes, and internal validity controls. Produces structured experiment plans conforming to experiments.schema.json. Works for any dimension where score_bruto is uncertain and evidence is missing. Neve

• idea-auditor-friction-analyst

  Use this agent when the friction dimension needs deep analysis — measuring how hard it is to reach first value (TTFV), activation rates, and where users drop off. Produces score_bruto (0–5) with evidence references conforming to evidence.schema.json. Never invents data; missing evidence results in s

• idea-auditor-loop-designer

  Use this agent when the loop dimension needs deep analysis — measuring referral mechanics, K-factor, and network effects. Produces score_bruto (0–5) with evidence references conforming to evidence.schema.json. Never invents data; missing evidence results in score_bruto=null. <example> Context: User

• idea-auditor-migration-analyst

  Use this agent when the migration dimension needs analysis — assessing switching costs, rollback feasibility, and diff surface for infrastructure forks or standards migrations. Active only in Infra_Fork_Standard mode. Produces score_bruto (0–5) with evidence references conforming to evidence.schema.

• idea-auditor-orchestrator

  Use this agent when running /idea-auditor:score or /idea-auditor:report to execute the full scoring pipeline in fixed sequence. Validates inputs, grades evidence from STATE/, scores all dimensions via deterministic scripts, and produces a PROCEED/ITERATE/KILL scorecard with top blockers and next tes

• idea-auditor-timing-scout

  Use this agent when the timing dimension needs deep analysis — identifying whether a window is open, acceleration signals are present, and catalysts can be named. Produces score_bruto (0–5) with evidence references conforming to evidence.schema.json. Never invents data; missing evidence results in s

• idea-auditor-trust-auditor

  Use this agent when the trust dimension needs deep analysis — measuring whether users take trust actions (give access, share data, accept risk) and whether the product's security posture supports that. Produces score_bruto (0–5) with evidence references conforming to evidence.schema.json. Never inve

• idea-auditor-wedge-researcher

  Use this agent when the wedge dimension needs deep analysis — assessing whether real pain exists, who feels it, and whether commitment signals are present. Produces score_bruto (0–5) with evidence references conforming to evidence.schema.json. Never invents data; missing evidence results in score_br

MCP servers (2)

• idea-auditor-evidence-harvester

  runs: python ${CLAUDE_PLUGIN_ROOT}/mcp/servers/evidence-harvester/server.py

• idea-auditor-analytics-bridge

  runs: python ${CLAUDE_PLUGIN_ROOT}/mcp/servers/analytics-bridge/server.py

Hooks (1)

• PostToolUse

  bash ${CLAUDE_PLUGIN_ROOT}/hooks/scripts/snapshot.sh

Similar plugins