Agent Recipes

Reusable workflows for open-source knowledge tasks.

Executable REST and MCP-oriented recipes for common open-source knowledge tasks: choose, compare, replace, inspect graph context, explore ecosystems, map Atlas journeys, check trust, and plan project stacks.

Open Recipes JSON Quickstart Agent Map Try Workflow

choose-cloudflare-agent-framework

Choose A Cloudflare-ready Agent Framework

JSON

A user wants an agent framework that fits Cloudflare or serverless deployment.

Outcome: Return a shortlist, explain why the top project fits, and cite deployment and confidence evidence.

GET/api/healthCheck live data
curl https://git.top/api/health
dbmetadata.sourcesync_freshness
GET/api/workflowBuild selection workflow
curl "https://git.top/api/workflow?intent=choose%20a%20Cloudflare-ready%20agent%20framework&deployment=cloudflare&category=agent_framework&cloudflare_ready=true&limit=5&require_d1=true"
recommended_sequenceshortlisttrust_policy
GET/api/project/cloudflare/agentsInspect top project
curl https://git.top/api/project/cloudflare/agents?require_d1=true
classificationquality_signal_confidencerelated
GET/api/compareCompare likely candidates
curl "https://git.top/api/compare?repos=cloudflare/agents,langchain-ai/langchain&deployment=cloudflare&require_d1=true"
decision_matrixwinnerreasoning
metadata.source=d1classification.cloudflare_ready.evidencerecommendations[].confidencequality_signal_confidence

find-project-alternatives

Find Alternatives To A Known Project

JSON

A user names a project and asks for replacements or adjacent options.

Outcome: Return direct substitutes and adjacent options with replacement risk and next comparison links.

GET/api/project/:projectResolve project
curl https://git.top/api/project/claude-code?require_d1=true
reporesolved_frommetadata.source
GET/api/alternatives/:projectFetch alternatives
curl https://git.top/api/alternatives/claude-code?limit=8&require_d1=true
alternative_matchesreplacement_riskcomparison_links
GET/api/compareCompare shortlist
curl "https://git.top/api/compare?repos=anthropics/claude-code,sst/opencode&deployment=local&require_d1=true"
decision_matrixtradeoffswinner
alternative_matches[].similarity_scorealternative_matches[].match_signalsreplacement_riskmetadata.source

compare-shortlist

Compare A Shortlist

JSON

A user already has two or more candidate repositories.

Outcome: Return a decision matrix and next actions instead of a star-only ranking.

GET/api/compareFetch comparison
curl "https://git.top/api/compare?repos=cloudflare/agents,langchain-ai/langchain,run-llama/llama_index&deployment=cloudflare&require_d1=true"
summarydecision_matrixwinnerreasoning
GET/api/score/:projectExplain score
curl https://git.top/api/score/cloudflare/agents?require_d1=true
dimensionsscore_confidencerisk_flags
GET/api/graph/:projectInspect graph context
curl https://git.top/api/graph/cloudflare/agents?limit=18&require_d1=true
relationship_groupsgraph_statsnext_actions
context.deploymentdecision_matrix[].tradeoffsscore_confidence.levelrelationship_groups

inspect-project-graph

Inspect Project Graph Context

JSON

A user wants to understand a project's alternatives, dependencies, related projects, and deployment targets.

Outcome: Return the graph summary, grouped relationships, and follow-up links.

GET/api/graph/:projectFetch graph
curl https://git.top/api/graph/browser-use/browser-use?limit=24&require_d1=true
summarygraph_statsrelationship_groupsnodesedges
GET/graph/:projectOpen human page
curl https://git.top/graph/browser-use/browser-use
Project overviewrelationship legendexploration paths
GET/api/alternatives/:projectFetch alternatives
curl https://git.top/api/alternatives/browser-use/browser-use?limit=6&require_d1=true
fit_summaryadoption_notesreplacement_risk
graph_stats.relationship_countsrelationship_groups.dependenciesmetadata.source

explore-ecosystem

Explore An Ecosystem

JSON

A user asks for a map of Cloudflare, MCP, RAG, AI agents, or Browser AI.

Outcome: Return representative projects, concept nodes, exploration paths, comparison paths, and next journeys.

GET/api/atlas/:ecosystemFetch Atlas
curl https://git.top/api/atlas/cloudflare?limit=8&require_d1=true
ecosystem.statsmap.nodesmap.edgesexploration_pathscomparison_paths
GET/atlas/:ecosystemBrowse human map
curl https://git.top/atlas/cloudflare
representative projectsjourneyscomparison pathsmap
GET/api/recommendRecommend within ecosystem
curl "https://git.top/api/recommend?use_case=build%20Cloudflare-ready%20AI%20agents&deployment=cloudflare&category=agent_framework&cloudflare_ready=true&require_d1=true"
recommendationsfit_profilerisk_flags
stats.project_countexploration_pathscomparison_paths[].contextrecommendations[].confidencemetadata.source

map-ecosystem-to-comparison

Map An Ecosystem To Comparison Paths

JSON

A user wants to move from an ecosystem map to concrete candidate comparisons.

Outcome: Return ecosystem-specific comparison paths, run the selected Compare JSON, and cite why the shortlist was chosen.

GET/api/journeysFetch journey index
curl https://git.top/api/journeys?limit=8&require_d1=true
journeyscomparison_pathsstats.comparison_path_countmetadata.source
GET/api/atlas/:ecosystemFetch ecosystem map
curl https://git.top/api/atlas/cloudflare?limit=8&require_d1=true
ecosystem.comparison_pathsecosystem.map.nodesecosystem.stats
GET/api/compareRun comparison path
curl "https://git.top/api/compare?repos=cloudflare/agents,langchain-ai/langchain&deployment=cloudflare&require_d1=true"
summarydecision_matrixwinnernext_actions
comparison_paths[].reposcomparison_paths[].decision_focusdecision_matrix[].tradeoffsmetadata.source=d1

check-recommendation-trust

Check Recommendation Trust

JSON

A user or agent needs to know whether recommendations are safe to cite.

Outcome: Return release score, data trust score, risk level, and disclosure guidance.

GET/api/healthFetch health
curl https://git.top/api/health
dbmetadata.sourcesync_healthsync_freshness
GET/api/qualityFetch quality report
curl https://git.top/api/quality?require_d1=true
release_scoredata_trust_scorerisk_levelimprovement_plancoverageissues
GET/api/quality/reviewFetch review queue
curl https://git.top/api/quality/review?require_d1=true
review_countitemsimpact_score
GET/coverageFetch coverage page
curl https://git.top/coverage
category distributioncollection semanticsuse boundaries
release_scoredata_trust_scorerisk_summary.reasonsimprovement_plancoverage.low_confidence_classification_rate

plan-with-grp

Plan With Graph Reasoning

JSON

A user asks for a plan, stack, or project set rather than a single repository.

Outcome: Return graph-grounded paths or stacks, then validate candidate projects through normal APIs.

POST/api/grp/queryRun GRP
curl -X POST "https://git.top/api/grp/query?require_d1=true" -H "content-type: application/json" -d '{"goal":"compose an autonomous coding stack with MCP tools","mode":"compose","constraints":{"agent_ready":true}}'
solution_pathsrecommended_stacknodesmetadata.data_source
GET/api/project/:projectInspect candidate
curl https://git.top/api/project/cloudflare/agents?require_d1=true
classificationrelatedquality_signal_confidence
GET/api/compareCompare final stack candidates
curl "https://git.top/api/compare?repos=cloudflare/agents,github/github-mcp-server&require_d1=true"
decision_matrixnext_actions
metadata.data_sourcenodes[].repoclassification evidencecomparison tradeoffs