Flow Bio Bridge
ClawBio's gateway to the Flow.bio platform — browse, search, upload, and launch bioinformatics pipelines on Flow from the command line.
Why This Exists
Flow.bio is a bioinformatics platform hosting curated Nextflow pipelines with managed compute, sample tracking, and collaborative project management. But interacting with Flow requires navigating the web UI or writing custom API scripts.
Flow Bio Bridge makes the platform agent-accessible: authenticate once, then list pipelines, upload samples, launch executions, and poll for results — all from the CLI or via the ClawBio orchestrator.
- Without it: Users must switch between web UI and local analysis, manually track execution IDs, and write one-off upload scripts
- With it: A single CLI covers discovery, upload, execution, and status checking across any Flow instance
- Why ClawBio: Chain Flow pipeline outputs with local ClawBio skills (e.g. Flow RNA-seq → ClawBio diffviz)
Core Capabilities
- Authentication — Login via username/password, existing JWT token, or environment variables
- Pipeline discovery — List and inspect available Nextflow pipelines with versions and parameter schemas
- Sample management — List, search, and upload samples with metadata and organism tagging
- Project browsing — List owned/shared projects and their contents
- Execution tracking — Launch pipeline runs, poll status, and retrieve logs
- Search — Full-text search across samples, projects, data, and executions
- Data listing — Browse owned and shared data files
- Organism & type discovery — List available organisms and sample types for uploads
- Overview mode — Live overview of any Flow instance; public data without credentials, full account view with login
Input Formats
| Format |
Extension |
Required Fields |
Example |
| FASTQ (single-end) |
.fq, .fastq, .fq.gz |
reads1 |
sample_R1.fastq.gz |
| FASTQ (paired-end) |
.fq, .fastq, .fq.gz |
reads1, reads2 |
sample_R1.fastq.gz, sample_R2.fastq.gz |
| Any data file |
any |
— |
annotation, reference, BED, etc. |
Workflow
When the user asks to interact with Flow.bio:
- Authenticate: Use stored token, env vars, or prompt for credentials
- Discover: List pipelines, sample types, organisms as needed
- Act: Upload samples, launch executions, or query status
- Report: Write structured output (JSON + markdown) to output directory
- Bundle: Generate reproducibility commands and checksums
CLI Reference
# Login (stores token for subsequent calls)
python skills/flow-bio/flow_bio.py --login --username USER --password PASS
python skills/flow-bio/flow_bio.py --login --token TOKEN
# Discovery
python skills/flow-bio/flow_bio.py --pipelines
python skills/flow-bio/flow_bio.py --samples
python skills/flow-bio/flow_bio.py --projects
python skills/flow-bio/flow_bio.py --organisms
python skills/flow-bio/flow_bio.py --sample-types
python skills/flow-bio/flow_bio.py --executions
python skills/flow-bio/flow_bio.py --data
# Inspect details
python skills/flow-bio/flow_bio.py --execution EXEC_ID
python skills/flow-bio/flow_bio.py --sample SAMPLE_ID
python skills/flow-bio/flow_bio.py --pipeline PIPELINE_ID
# Inspect details via clawbio.py runner (uses --*-detail flags)
python clawbio.py run flow --execution-detail EXEC_ID
python clawbio.py run flow --sample-detail SAMPLE_ID
python clawbio.py run flow --pipeline-detail PIPELINE_ID
# Search
python skills/flow-bio/flow_bio.py --search "RNA-seq tumor"
# Raw JSON output (for piping to jq, etc.)
python skills/flow-bio/flow_bio.py --samples --json
# Upload sample
python skills/flow-bio/flow_bio.py --upload-sample \
--name "Tumour_01" --sample-type "RNA-Seq" \
--reads1 R1.fastq.gz --reads2 R2.fastq.gz \
--organism "Homo sapiens" --project PROJECT_ID
# Launch pipeline
python skills/flow-bio/flow_bio.py --run-pipeline PIPELINE_VERSION_ID \
--run-samples SAMPLE_ID1,SAMPLE_ID2 --output /tmp/flow_run
# Check execution status
python skills/flow-bio/flow_bio.py --execution EXEC_ID --output /tmp/flow_status
# Overview (public endpoints, no credentials needed)
python skills/flow-bio/flow_bio.py --demo --output /tmp/flow_demo
# Overview with authentication (shows owned samples, projects, executions)
python skills/flow-bio/flow_bio.py --demo --username USER --password PASS --output /tmp/flow_demo
Demo
# Public overview (no credentials)
python skills/flow-bio/flow_bio.py --demo --output /tmp/flow_demo
# Full overview with account data
python skills/flow-bio/flow_bio.py --demo --username USER --password PASS --output /tmp/flow_demo
Expected output: a live overview of the Flow.bio instance showing available
pipelines, organisms, and sample types (public), plus owned samples, projects,
and executions if authenticated.
Algorithm / Methodology
- Token acquisition: POST
/login with credentials → JWT token (5-min access, 7-day refresh)
- API traversal: RESTful GET/POST with
Authorization: Bearer <token> header
- Pagination: Offset-based (
?page=N) with lazy loading
- Chunked upload: Files split into 1 MB chunks, uploaded sequentially with progress tracking
- Execution polling: GET
/executions/<id> until status reaches terminal state
Key parameters:
- Base URL:
https://app.flow.bio/api (configurable via FLOW_URL)
- Chunk size: 1 MB (configurable)
- Rate limits: 1000 req/hr API, 100/hr uploads, 500/hr downloads
Example Queries
- "What pipelines are available on Flow?"
- "List my samples on Flow.bio"
- "Upload these FASTQ files to Flow"
- "Run the RNA-seq pipeline on my sample"
- "What's the status of my Flow execution?"
- "Search Flow for breast cancer samples"
Output Structure
output_dir/
├── report.md # Summary of Flow API interaction
├── result.json # Machine-readable response data
└── reproducibility/
├── commands.sh # Portable replay recipe (CLAWBIO_ROOT / OUTPUT_DIR)
├── environment.yml # Conda environment for the run (python + requests)
└── checksums.sha256 # SHA-256 of report.md and result.json
Dependencies
Required:
requests >= 2.28 — HTTP client for REST API calls
Safety
- Credentials never stored on disk — tokens held in memory or via environment variables only
- No destructive defaults — delete/transfer operations are not exposed in v0.1
- Audit trail: Every API interaction logged to reproducibility bundle
- Disclaimer: ClawBio is a research and educational tool. It is not a medical device and does not provide clinical diagnoses. Consult a healthcare professional before making any medical decisions.
Integration with Bio Orchestrator
Trigger conditions — the orchestrator routes here when:
- User mentions "flow", "flow.bio", "flow pipeline", "run on flow"
- User wants to upload samples to a cloud platform
- User asks about pipeline execution status on Flow
Chaining partners:
rnaseq-de — Flow RNA-seq pipeline output → ClawBio differential expression
diffviz — Flow DE results → ClawBio visualization
scrna-orchestrator — Flow Cell Ranger output → ClawBio scRNA analysis
illumina-bridge — Illumina DRAGEN → Flow upload → pipeline execution
Citations
- Flow.bio — Bioinformatics platform
- Nextflow — Workflow engine powering Flow pipelines
1---2name: flow-bio3description: Flow.bio API bridge — authenticate, browse pipelines/samples/projects, search, upload data, launch pipeline executions,4 and check run status on any Flow instance.5license: MIT6metadata:7 source: https://skillmd.com/skills/clawbio/flow-bio8 version: 0.1.09 author: ClawBio Contributors10 domain: bioinformatics11 tags:12 - flow13 - flow.bio14 - LIMS15 - pipelines16 - samples17 - nextflow18 - bioinformatics19 - cloud20 inputs:21 - name: reads122 type: file23 format:24 - fastq25 - fq26 - fastq.gz27 - fq.gz28 description: First reads file for sample upload29 required: false30 - name: reads231 type: file32 format:33 - fastq34 - fq35 - fastq.gz36 - fq.gz37 description: Second reads file for paired-end upload38 required: false39 outputs:40 - name: report41 type: file42 format: md43 description: Markdown report of Flow.bio interaction44 - name: result45 type: file46 format: json47 description: Machine-readable result envelope48 dependencies:49 python: '>=3.10'50 packages:51 - requests>=2.2852 demo_data:53 - path: data/demo_cache.json54 description: Pre-cached public Flow.bio data (pipelines, organisms, sample types) for offline demo55 endpoints:56 cli: python skills/flow-bio/flow_bio.py --demo --output {output_dir}57 openclaw:58 requires:59 bins:60 - python361 env:62 - FLOW_URL63 - FLOW_TOKEN64 always: false65 emoji: 🌊66 homepage: https://flow.bio67 os:68 - darwin69 - linux70 install:71 - kind: pip72 package: requests73 trigger_keywords:74 - flow75 - flow.bio76 - flow bio77 - flow pipeline78 - flow sample79 - flow execution80 - flow project81 - flow upload82 - run on flow83---8485# Flow Bio Bridge8687**ClawBio's gateway to the Flow.bio platform — browse, search, upload, and launch bioinformatics pipelines on Flow from the command line.**8889## Why This Exists9091Flow.bio is a bioinformatics platform hosting curated Nextflow pipelines with managed compute, sample tracking, and collaborative project management. But interacting with Flow requires navigating the web UI or writing custom API scripts.9293Flow Bio Bridge makes the platform **agent-accessible**: authenticate once, then list pipelines, upload samples, launch executions, and poll for results — all from the CLI or via the ClawBio orchestrator.9495- **Without it**: Users must switch between web UI and local analysis, manually track execution IDs, and write one-off upload scripts96- **With it**: A single CLI covers discovery, upload, execution, and status checking across any Flow instance97- **Why ClawBio**: Chain Flow pipeline outputs with local ClawBio skills (e.g. Flow RNA-seq → ClawBio diffviz)9899## Core Capabilities1001011. **Authentication** — Login via username/password, existing JWT token, or environment variables1022. **Pipeline discovery** — List and inspect available Nextflow pipelines with versions and parameter schemas1033. **Sample management** — List, search, and upload samples with metadata and organism tagging1044. **Project browsing** — List owned/shared projects and their contents1055. **Execution tracking** — Launch pipeline runs, poll status, and retrieve logs1066. **Search** — Full-text search across samples, projects, data, and executions1077. **Data listing** — Browse owned and shared data files1088. **Organism & type discovery** — List available organisms and sample types for uploads1099. **Overview mode** — Live overview of any Flow instance; public data without credentials, full account view with login110111## Input Formats112113| Format | Extension | Required Fields | Example |114|--------|-----------|-----------------|---------|115| FASTQ (single-end) | `.fq`, `.fastq`, `.fq.gz` | reads1 | `sample_R1.fastq.gz` |116| FASTQ (paired-end) | `.fq`, `.fastq`, `.fq.gz` | reads1, reads2 | `sample_R1.fastq.gz`, `sample_R2.fastq.gz` |117| Any data file | any | — | annotation, reference, BED, etc. |118119## Workflow120121When the user asks to interact with Flow.bio:1221231. **Authenticate**: Use stored token, env vars, or prompt for credentials1242. **Discover**: List pipelines, sample types, organisms as needed1253. **Act**: Upload samples, launch executions, or query status1264. **Report**: Write structured output (JSON + markdown) to output directory1275. **Bundle**: Generate reproducibility commands and checksums128129## CLI Reference130131```bash132# Login (stores token for subsequent calls)133python skills/flow-bio/flow_bio.py --login --username USER --password PASS134python skills/flow-bio/flow_bio.py --login --token TOKEN135136# Discovery137python skills/flow-bio/flow_bio.py --pipelines138python skills/flow-bio/flow_bio.py --samples139python skills/flow-bio/flow_bio.py --projects140python skills/flow-bio/flow_bio.py --organisms141python skills/flow-bio/flow_bio.py --sample-types142python skills/flow-bio/flow_bio.py --executions143python skills/flow-bio/flow_bio.py --data144145# Inspect details146python skills/flow-bio/flow_bio.py --execution EXEC_ID147python skills/flow-bio/flow_bio.py --sample SAMPLE_ID148python skills/flow-bio/flow_bio.py --pipeline PIPELINE_ID149150# Inspect details via clawbio.py runner (uses --*-detail flags)151python clawbio.py run flow --execution-detail EXEC_ID152python clawbio.py run flow --sample-detail SAMPLE_ID153python clawbio.py run flow --pipeline-detail PIPELINE_ID154155# Search156python skills/flow-bio/flow_bio.py --search "RNA-seq tumor"157158# Raw JSON output (for piping to jq, etc.)159python skills/flow-bio/flow_bio.py --samples --json160161# Upload sample162python skills/flow-bio/flow_bio.py --upload-sample \163 --name "Tumour_01" --sample-type "RNA-Seq" \164 --reads1 R1.fastq.gz --reads2 R2.fastq.gz \165 --organism "Homo sapiens" --project PROJECT_ID166167# Launch pipeline168python skills/flow-bio/flow_bio.py --run-pipeline PIPELINE_VERSION_ID \169 --run-samples SAMPLE_ID1,SAMPLE_ID2 --output /tmp/flow_run170171# Check execution status172python skills/flow-bio/flow_bio.py --execution EXEC_ID --output /tmp/flow_status173174# Overview (public endpoints, no credentials needed)175python skills/flow-bio/flow_bio.py --demo --output /tmp/flow_demo176177# Overview with authentication (shows owned samples, projects, executions)178python skills/flow-bio/flow_bio.py --demo --username USER --password PASS --output /tmp/flow_demo179```180181## Demo182183```bash184# Public overview (no credentials)185python skills/flow-bio/flow_bio.py --demo --output /tmp/flow_demo186187# Full overview with account data188python skills/flow-bio/flow_bio.py --demo --username USER --password PASS --output /tmp/flow_demo189```190191Expected output: a live overview of the Flow.bio instance showing available192pipelines, organisms, and sample types (public), plus owned samples, projects,193and executions if authenticated.194195## Algorithm / Methodology1961971. **Token acquisition**: POST `/login` with credentials → JWT token (5-min access, 7-day refresh)1982. **API traversal**: RESTful GET/POST with `Authorization: Bearer <token>` header1993. **Pagination**: Offset-based (`?page=N`) with lazy loading2004. **Chunked upload**: Files split into 1 MB chunks, uploaded sequentially with progress tracking2015. **Execution polling**: GET `/executions/<id>` until status reaches terminal state202203**Key parameters**:204- Base URL: `https://app.flow.bio/api` (configurable via `FLOW_URL`)205- Chunk size: 1 MB (configurable)206- Rate limits: 1000 req/hr API, 100/hr uploads, 500/hr downloads207208## Example Queries209210- "What pipelines are available on Flow?"211- "List my samples on Flow.bio"212- "Upload these FASTQ files to Flow"213- "Run the RNA-seq pipeline on my sample"214- "What's the status of my Flow execution?"215- "Search Flow for breast cancer samples"216217## Output Structure218219```220output_dir/221├── report.md # Summary of Flow API interaction222├── result.json # Machine-readable response data223└── reproducibility/224 ├── commands.sh # Portable replay recipe (CLAWBIO_ROOT / OUTPUT_DIR)225 ├── environment.yml # Conda environment for the run (python + requests)226 └── checksums.sha256 # SHA-256 of report.md and result.json227```228229## Dependencies230231**Required:**232- `requests` >= 2.28 — HTTP client for REST API calls233234## Safety235236- **Credentials never stored on disk** — tokens held in memory or via environment variables only237- **No destructive defaults** — delete/transfer operations are not exposed in v0.1238- **Audit trail**: Every API interaction logged to reproducibility bundle239- **Disclaimer**: ClawBio is a research and educational tool. It is not a medical device and does not provide clinical diagnoses. Consult a healthcare professional before making any medical decisions.240241## Integration with Bio Orchestrator242243**Trigger conditions** — the orchestrator routes here when:244- User mentions "flow", "flow.bio", "flow pipeline", "run on flow"245- User wants to upload samples to a cloud platform246- User asks about pipeline execution status on Flow247248**Chaining partners**:249- `rnaseq-de` — Flow RNA-seq pipeline output → ClawBio differential expression250- `diffviz` — Flow DE results → ClawBio visualization251- `scrna-orchestrator` — Flow Cell Ranger output → ClawBio scRNA analysis252- `illumina-bridge` — Illumina DRAGEN → Flow upload → pipeline execution253254## Citations255256- [Flow.bio](https://flow.bio) — Bioinformatics platform257- [Nextflow](https://nextflow.io/) — Workflow engine powering Flow pipelines