About artifactbloat
artifactbloat is a local CLI and MCP server for tracking GitHub Actions artifact sizes by workflow. It compares dated snapshots, reports growth and flags workflows that cross a configurable threshold.
Key features
- 01Per-workflow snapshot totals
- 02First-to-last growth deltas
- 03Configurable growth flags
- 04Per-artifact size comparisons
- 05Overall byte totals
Use cases
- 01Spot workflows with rapidly growing artifacts
- 02Identify when artifacts appeared or disappeared
- 03Review size changes across dated snapshots
MCP support
Exposes a `track_artifact_sizes` tool that accepts dated workflow snapshots and an optional growth threshold. It returns totals, workflow details, flagged workflows, and a verdict.
Configuration
Local server config is the same for every client. Copy it into your client’s MCP config file.
Tools
1Saved tool list · updated 1w ago
| Parameter | Type | Required | Description |
|---|---|---|---|
| snapshots | array | Yes | Artifact snapshots to track (at least one) |
| growth_threshold_percent | number | No | Growth threshold in percent for flagging a workflow (optional, default 20) |
Details
From the project READMEartifactbloat
Your CI artifacts quietly grow a few percent every week — until a workflow that used to upload a slim bundle is shipping something enormous and nobody can say when it happened. artifactbloat takes dated snapshots of artifact sizes, groups them per workflow, and shows the growth in plain numbers:
- Per-workflow totals — snapshots ordered by date, with the first, last, and largest snapshot totals for every workflow
- Growth deltas — the change from first to last
snapshot in bytes and as a percentage (rounded to four
decimal places;
nullwhen the first total is zero, so a workflow that started empty is never divided by zero) - Bloat flags — a workflow is flagged when its growth
meets a threshold you choose (20 percent by default),
boundary included; the flagged names are listed on their
own and drive the verdict:
growth_flaggedwhen anything is flagged,steadyotherwise - Per-artifact detail — for every artifact name a
workflow ever produced: its size in the first snapshot
and in the last, with
nullon either side when the artifact did not exist there (a new artifact, a removed one, or one seen only in between), and the byte delta when both sides exist - Overall totals — the summed first and last totals across all workflows
The computation is pure and local: no network, no configuration, no files read beyond the snapshots you pass in. Invalid input — a missing snapshot list, a snapshot without a workflow, a negative size, the same workflow captured twice on one date, or the same artifact twice in one snapshot — is rejected with an error naming the offending path, and no partial report is produced.
CLI
npx @reneturcios/artifactbloat track --input args.json
npx @reneturcios/artifactbloat --version
args.json contains the full arguments object:
{
"snapshots": [
{
"workflow": "build",
"date": "2026-01-01",
"artifacts": [
{ "name": "app-bundle", "size_bytes": 1048576 },
{ "name": "source-maps", "size_bytes": 262144 }
]
},
{
"workflow": "build",
"date": "2026-02-01",
"artifacts": [
{ "name": "app-bundle", "size_bytes": 1572864 },
{ "name": "source-maps", "size_bytes": 262144 }
]
}
],
"growth_threshold_percent": 20
}
growth_threshold_percent is optional and defaults to
20. The report is printed as JSON.
Exit codes: 0 on success, 1 on a runtime error
(including an unreadable or invalid input file, or invalid
input), 2 on a usage error.
MCP server
npx -y @reneturcios/artifactbloat mcp
MCP client config:
{
"mcpServers": {
"artifactbloat": {
"command": "npx",
"args": ["-y", "@reneturcios/artifactbloat", "mcp"]
}
}
}
Tools
-
track_artifact_sizes— Track Actions artifact sizes over time, per workflow.snapshots(array, required, at least one) — snapshots as{workflow, date, artifacts}objects, where each artifact is{name, size_bytes}anddateis the capture day,YYYY-MM-DDgrowth_threshold_percent(number, optional, >= 0, default 20) — flag workflows whose total grew by at least this percent from first to last snapshot
Returns
{workflow_count, total_first_bytes, total_last_bytes, workflows, flagged_workflows, verdict}.
Docker
docker run -i ghcr.io/cloudcorprecords/artifactbloat
docker run --rm -v "$PWD:/w" ghcr.io/cloudcorprecords/artifactbloat track --input /w/args.json
Development
npm install
npm run build # tsc -> dist/
npm test # tsx --test test/*.test.ts (pure local, no network)
npm run dev # run the CLI from source
License
MIT — Copyright (c) 2026 Rene Turcios
Comments
New to MCP?
From what the protocol is to installing your first server, these answers cover most first questions.
An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.