skills/forge-compress/SKILL.md
[writes] Unified compression — `agents` compresses agent .md files for 30-50% system-prompt reduction; `output <mode>` sets runtime output compression (off|lite|full|ultra) writing .forge/caveman-mode; `status` shows current settings (default, read-only); `help` prints reference card. Use when you want to save tokens on prompts or session output. Trigger: /forge-compress, compress agents, compress output, caveman mode, reduce tokens
npx skillsauth add quantumbitcz/dev-pipeline forge-compressInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
Single entry point for compression. Replaces /forge-compress (previous agent-only surface), /forge-caveman, and /forge-compression-help (all removed in 3.0.0).
| Subcommand | Read/Write | Purpose |
|---|---|---|
| agents | writes | Compress agent .md files via terse-rewrite (30–50% reduction) |
| output <mode> | writes | Set output compression. mode ∈ {off, lite, full, ultra}. Writes .forge/caveman-mode |
| status (default) | read-only | Show current agent-compression ratio and output-mode |
| help | read-only | Reference card (flags, modes, token savings by mode, tips) |
See shared/skill-contract.md for the standard exit-code table.
agents/ directory present (for agents subcommand).forge/ directory writable (for output subcommand, which persists .forge/caveman-mode)Dispatch the compression op to the orchestrator or the underlying compression pipeline:
agents: compress all agents/fg-*.md files via terse rewriting; preserves code blocks, YAML frontmatter, and all technical rules. See shared/output-compression.md and shared/agent-ui.md.output <mode>: write the mode string (off|lite|full|ultra) to .forge/caveman-mode. The runtime reads this at agent dispatch time to select the per-stage compression level.status (default): read current agent file sizes and .forge/caveman-mode, report ratios.help: emit the reference card inline.Exit codes per shared/skill-contract.md:
output)/forge-insights -- View compression-effectiveness analytics across runs/forge-config -- Toggle compression defaults in forge-config.md/forge-status -- Check current compression mode at runtimeshared/skill-contract.md — standard exit codes and flag conventionsshared/output-compression.md — per-stage output compression level modelshared/input-compression.md — agent prompt compression pipeline/forge-compress # default: status
/forge-compress output lite # set lite mode
/forge-compress output ultra --dry-run # preview ultra without writing
/forge-compress agents # compress all agent .md
/forge-compress agents --dry-run # preview compression
/forge-compress help # reference card
/forge-compress status --json # JSON for scripting
| Mode | Token savings | Description | |------|---------------|-------------| | off | 0% | Full verbose output (default) | | lite | ~30% | Strip redundant narration; keep code/data intact | | full | ~55% | Aggressive prose compression; ellipsis-heavy | | ultra | ~75% | Caveman grammar; skeletal output only |
Replacements for removed skills:
| Old skill | New invocation | |---|---| | /forge-caveman | /forge-compress output <mode> | | /forge-compression-help | /forge-compress help |
development
[writes] Build, fix, deploy, review, or modify code in this project. Universal entry for the forge pipeline. Auto-bootstraps on first run; brainstorms before planning when given a feature description. Use when you want to take any productive action: implementing features, fixing bugs, reviewing branches, deploying, committing, running migrations.
tools
[writes] Manage forge state and configuration: recovery, abort, config edits, session handoff, automations, playbooks, output compression, knowledge graph maintenance. Use when you need to recover from broken pipeline state, edit settings, or manage long-lived state.
development
[writes] Create, list, show, resume, or search forge session handoffs. Use when context is getting heavy and you want to transfer a forge run or conversation into a fresh Claude Code session, or to resume from a prior handoff artefact. Subcommands - no args (write), list, show, resume, search.
development
[writes] Manage the Neo4j knowledge graph. Subcommands: init, rebuild (writes); status, query <cypher>, debug (read-only). Requires Docker. No default — an explicit subcommand is required. Use when setting up the graph for the first time, rebuilding after major refactors, checking graph health, or running ad-hoc Cypher diagnostics.