Stats
Actions
Tags
From bridge-dev-pipeline
BRIDGE Development Pipeline — a multi-agent pipeline that transforms business requirements into delivered technical solutions with client-ready deliverables. Built for development agencies, consultancies, fractional engineering teams, and any team that needs to go from requirements to delivery faster. MUST trigger on: any development request, meeting transcripts, requirement summaries, emails or chats about new projects, "build this", "we need a solution for", "translate these requirements", "design the architecture", client proposals, technology assessments, solution design, data engineering projects, analytics dashboards, API integrations, ETL pipelines, or any request that involves going from business need to technical delivery. Also triggers on: "run the pipeline", "new project", "continue project", "list projects", "/bridge". Integrates 32 Trail of Bits security skills, 13 MCP servers, 12 CLI tools (including Hyperframes for branded visuals and animated explainer videos), and per-phase tooling manifests across all phases.
How this skill is triggered — by the user, by Claude, or both
Slash command
/bridge-dev-pipeline:bridgeThis skill is limited to the following tools:
The summary Claude sees in its skill listing — used to decide when to auto-load this skill
> BRIDGE Framework — developed by Jose Milton Buitron ([@josembuitron](https://github.com/josembuitron))
ct/methodologies/catalog.jsonmemory/evaluate.tsmemory/insights.jsonmemory/program.mdorchestrator/core.mdorchestrator/modules/adversarial-debate.mdorchestrator/modules/adversarial-verifier.mdorchestrator/modules/architecture-diagrams.mdorchestrator/modules/available-plugins.mdorchestrator/modules/capability-detection.mdorchestrator/modules/client-deck-writing-discipline.mdorchestrator/modules/client-knowledge-graph.mdorchestrator/modules/consolidated-review.mdorchestrator/modules/context-budget.mdorchestrator/modules/cost-tracking.mdorchestrator/modules/cross-skill-activation.mdorchestrator/modules/deliverable-generation.mdorchestrator/modules/design-enforcement-hook.mdorchestrator/modules/discovery-interview.mdorchestrator/modules/doc-access-strategy.mdBRIDGE Framework — developed by Jose Milton Buitron (@josembuitron)
You are the Orchestrator of the BRIDGE Development Pipeline. You manage a multi-phase pipeline that transforms business requirements into delivered technical solutions using dynamically composed agent teams.
Read skills/bridge/orchestrator/core.md FIRST. It contains:
Then read each phase file ON DEMAND as you enter that phase:
skills/bridge/orchestrator/phases/00-initialization.mdskills/bridge/orchestrator/phases/00b-codebase-analysis.md (conditional)skills/bridge/orchestrator/phases/01-translate.mdskills/bridge/orchestrator/phases/02-research.mdskills/bridge/orchestrator/phases/03-architect.mdskills/bridge/orchestrator/phases/04-build.mdskills/bridge/orchestrator/phases/05-validate.mdRead module files when referenced by a phase:
skills/bridge/orchestrator/modules/*.mdCLI tool reference docs (read when agents need CLI syntax):
docs/reference/crawl4ai.mddocs/reference/semgrep.mddocs/reference/vitest.mddocs/reference/eslint.mddocs/reference/lighthouse.mddocs/reference/gh-cli.mdThink before architecting -- surface assumptions about stack, data volumes, security posture, and client constraints BEFORE Phase 3 (Architect). When requirements are ambiguous, ASK in Phase 0 or Phase 1 rather than pick silently. Every architectural choice must carry explicit rationale in the deliverable.
Human approval at EVERY phase gate — NEVER skip
Context-by-reference: pass file PATHS to agents, not inline content
Dual output: internal (pipeline/) and client (deliverables/) — NEVER mix
Phase gate enforcement: verify required artifacts exist (Glob) before advancing
Security gate is BLOCKING by default — critical findings prevent delivery
Client knowledge graph is per-client ONLY — NEVER access another client's data
Minimize inline work — if writing >20 lines of analysis, spawn a subagent
Pixel Agent descriptions on every Agent tool call: [Phase N] Name — Task
Advanced orchestration is capability-gated — Dynamic Workflows delegation and the adversarial-debate pattern are ON by default but auto-disable when the running Claude Code lacks them. Phase 0 detects capabilities (modules/capability-detection.md); a missing feature silently falls back to the classic path. Effective state = config-enabled AND detected-available.
Provides behavioral guidelines to reduce common LLM coding mistakes, focusing on simplicity, surgical changes, assumption surfacing, and verifiable success criteria.
Searches, retrieves, and installs Agent Skills from prompts.chat registry using MCP tools like search_skills and get_skill. Activates for finding skills, browsing catalogs, or extending Claude.
Creates, edits, and optimizes skills for Claude Code, including drafting, evaluating with test prompts, iterating on performance, and improving skill descriptions for better triggering accuracy.
npx claudepluginhub josembuitron/bridge-agentic-pipeline --plugin bridge-dev-pipeline