How do I install the Batch Compress skill?

Run npx skillmds@latest add oimiragieo/batch-compress in your terminal (requires Node.js), paste this page's agent-chat prompt into Claude, Cursor, or any MCP-connected agent, or download the SKILL.md file and copy it into your agent's skills directory.

Is Batch Compress safe to use?

This skill has not completed SkillMD's automated safety review yet. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.

Which AI agents work with Batch Compress?

This skill is tagged as working with Claude Code, Claude.ai, OpenAI Codex. SKILL.md is an open format, so most agents that read a skills directory can load it too.

Is Batch Compress free to use?

Yes. Installing skills from SkillMD is free, and the skill stays under its author's original license.

Who published Batch Compress?

oimiragieo (@oimiragieo) published this skill. Their other Agent Skills are listed on their SkillMD profile.

Batch Compress by oimiragieo — Agent Skill

Name: Batch Compress
Author: oimiragieo

Batch Compress

Bulk compression of a file corpus — up to 50 documents per call — for later queries rather than one-off reads. Use this skill whenever the user wants to pre-process a directory, a support-ticket dump, meeting notes, or any sizable corpus before Q&A. Trigger phrases include "compress all these files", "pre-process this corpus", "shrink everything in /path", "bulk-compress for Q&A", or when onboarding Claude to a large project that needs a lot of ingested context. Submits an async job to gotcontext's batch queue and returns a job id; the job processes in the background with per-item error reporting so one bad file doesn't block the batch. Pro, Team, or Enterprise plan required.

oimiragieo faa1e06 2.3 KB Updated Sep 17, 2026 0 repo stars

File contents

batch-compress

When this skill is the preferred path

For corpus-scale pre-processing, batching amortizes network round-trips and returns one aggregate savings report. A loop of ingest_context calls is the slow and expensive path for the same work.

How to use it

Enumerate the files. Respect .gitignore. Skip binaries, generated files, lockfiles, node_modules, vendor, dist, build.
Group into batches of up to 50 documents (the API cap). Larger corpora need multiple calls.
Call batch_ingest_documents (MCP tool; Pro+). Returns a job_id immediately.
Poll job status via GET /v1/batch-queue/{id} (REST) or subscribe to GET /v1/batch-queue/stream for SSE updates.
When complete, retrieve each compressed item by its item id.
Report to the user:
- Total files processed
- Aggregate tokens saved
- Any failures (rate limit, invalid input, permissions) — the API returns per-item errors so failures don't block the batch

When the raw read or a simpler skill fits better

Single file or a small set (<3 files) — shrink-for-claude.
Files the user is actively editing — stale snapshots will confuse.

Plan gating

batch_ingest_documents returns HTTP 403 on Free with an upgrade link. Surface that response faithfully — don't paper over it with a fallback loop. The user wanted batch throughput; they should see the upgrade path instead of a silent slow fallback.

Warn before starting

If the batch will have >200 items, confirm before enqueuing. Processing time scales linearly and the user may have wanted a smaller slice.

---name: batch-compressdescription: Bulk compression of a file corpus — up to 50 documents per call — for later queries rather than one-off reads. Use this skill whenever the user wants to pre-process a directory, a support-ticket dump, meeting notes, or any sizable corpus before Q&A. Trigger phrases include "compress all these files", "pre-process this corpus", "shrink everything in /path", "bulk-compress for Q&A", or when onboarding Claude to a large project that needs a lot of ingested context. Submits an async job to gotcontext's batch queue and returns a job id; the job processes in the background with per-item error reporting so one bad file doesn't block the batch. Pro, Team, or Enterprise plan required.---# batch-compress## When this skill is the preferred pathFor corpus-scale pre-processing, batching amortizes network round-tripsand returns one aggregate savings report. A loop of `ingest_context`calls is the slow and expensive path for the same work.## How to use it1. Enumerate the files. Respect `.gitignore`. Skip binaries, generated files, lockfiles, `node_modules`, `vendor`, `dist`, `build`.2. Group into batches of up to 50 documents (the API cap). Larger corpora need multiple calls.3. Call `batch_ingest_documents` (MCP tool; Pro+). Returns a `job_id` immediately.4. Poll job status via `GET /v1/batch-queue/{id}` (REST) or subscribe to `GET /v1/batch-queue/stream` for SSE updates.5. When complete, retrieve each compressed item by its item id.6. Report to the user: - Total files processed - Aggregate tokens saved - Any failures (rate limit, invalid input, permissions) — the API returns per-item errors so failures don't block the batch## When the raw read or a simpler skill fits better- Single file or a small set (<3 files) — `shrink-for-claude`.- Files the user is actively editing — stale snapshots will confuse.## Plan gating`batch_ingest_documents` returns HTTP 403 on Free with an upgradelink. Surface that response faithfully — don't paper over it with afallback loop. The user wanted batch throughput; they should see theupgrade path instead of a silent slow fallback.## Warn before startingIf the batch will have >200 items, confirm before enqueuing.Processing time scales linearly and the user may have wanted a smallerslice.

Frequently asked questions