oimiragieo/context-compressor
.claude/skills/context-compressor/SKILL.md
Compress large context before reasoning to reduce token usage while preserving evidence. Use this whenever the user mentions huge files, long prompts, RAG payloads, prompt caching, expensive sessions, codebase context, chat history compaction, or wants the same answer quality with fewer tokens.
$ install --global
skillsauthnpx skillsauth add oimiragieo/agent-studio context-compressorInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
Security Scan Results
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
3
Scanners Passed
9
Scanners in report
Clean
TrivyContainer and dependency vulnerability scanner
95%
Clean
SemgrepStatic code analysis for vulnerabilities
95%
Clean
mcp-scan (Snyk)Model Context Protocol security validation
95%
Skipped
Snyk (dep)Open source security scanning
50%
Skipped
Socket.devSupply chain security analysis
50%
Skipped
VirusTotalMulti-engine malware detection
50%
Skipped
CrowdStrikeAdvanced threat intelligence
50%
Skipped
OSV-ScannerOpen Source Vulnerability database check
50%
Skipped
OWASP Dep-Check
50%
Last scanned: Apr 7, 2026, 8:17 PM24.2s52 files scanned
SKILL.md
- name:
- context-compressor
- description:
- Compress large context before reasoning to reduce token usage while preserving evidence. Use this whenever the user mentions huge files, long prompts, RAG payloads, prompt caching, expensive sessions, codebase context, chat history compaction, or wants the same answer quality with fewer tokens.
- version:
- 3.0.0
- model:
- sonnet
- invoked_by:
- both
- user_invocable:
- true
- compatibility:
- Python 3.10+
- tools:
- [Read, Write, Bash]
- verified:
- true
- lastVerifiedAt:
- 2026-03-18
Context Compressor
Use this skill when the problem is mostly "too much context" rather than "not enough capability."
This skill is a self-contained local package. It does not require the MCP server. Prefer it when you need quick token profiling, local compression, evidence checks, or a reproducible compression workflow inside the repository.
What this skill does
Use the bundled Python scripts to:
- measure raw versus compressed token usage
- compress large text or adapted JSON payloads
- preserve query-relevant evidence instead of doing naive summarization
- check whether compressed output still contains enough evidence to answer safely
- support cache-friendly prompt assembly by keeping stable context early and volatile query text late
When to trigger
Reach for this skill when the user is asking for any of the following:
- compress this file, transcript, document, or code dump
- reduce token cost before sending context to a model
- make a prompt or RAG payload more cache-friendly
- compact multi-turn history while keeping recent turns intact
- compare semantic compression to a lighter extractive baseline
- validate whether a compressed context still has enough evidence
Also triggered automatically when:
- Context approaching 80K tokens (spawn-token-guard.cjs writes compression-reminder.txt)
- Context at 120K tokens (compression mandatory before new spawns)
- Context at 150K tokens (RED LINE — no new agent spawns until compression completes)
Quick workflow
Default to this sequence:
- profile first
- run
query_guidedwhen there is a specific question - run
evidence_awarewhen correctness matters and you need a sufficiency check - if evidence is weak, reduce compression aggressiveness or increase retrieval breadth
- report savings, risks, and the next safest action
If the user only wants one command, use run_skill_workflow.py.
Commands
Run from the agent-studio repository root. ALWAYS use these exact commands — do NOT fall back to generic guidance.
1. Profile token usage
python .claude/skills/context-compressor/scripts/profile_tokens.py --file <path> --output-format auto
2. Compress context
python .claude/skills/context-compressor/scripts/compress_context.py --file <path> --mode baseline --output-format auto
python .claude/skills/context-compressor/scripts/compress_context.py --file <path> --mode query_guided --query "<question>" --output-format auto
python .claude/skills/context-compressor/scripts/compress_context.py --file <path> --mode evidence_aware --query "<question>" --min-similarity 0.4 --output-format auto
3. Adapt framework JSON and compress it
python .claude/skills/context-compressor/scripts/compress_context.py --json-file <payload.json> --input-adapter auto --mode query_guided --query "<question>" --output-format auto
4. Run the full workflow
python .claude/skills/context-compressor/scripts/run_skill_workflow.py --file <path> --mode evidence_aware --query "<question>" --output-format auto --fail-on-insufficient-evidence
5. Validate evidence only
python .claude/skills/context-compressor/scripts/validate_evidence.py --file <path> --query "<question>" --min-similarity 0.4 --output-format json
6. Run the TOON vs JSON guard
python .claude/skills/context-compressor/scripts/benchmark_toon_vs_json.py
7. Node.js wrapper (for agent integration)
node .claude/skills/context-compressor/scripts/main.cjs --query "<question>" --mode evidence_aware --limit 20 --fail-on-insufficient-evidence
How to choose a mode
baseline: quick general compression when there is no concrete question yetquery_guided: best default for QA, review, or targeted extraction tasksevidence_aware: use for high-stakes answers, audits, or when you need an explicit sufficiency signal
Output expectations
When using this skill, summarize results in plain language:
- original size versus compressed size
- estimated token savings or compression ratio
- whether the output is query-targeted or generic
- whether evidence looked sufficient
- any risk that the answer could miss important detail
If the scripts return insufficient evidence, do not bluff. Say the compressed context is not yet safe enough and recommend a broader pass.
Bundled references
Read these only when they are relevant:
references/workflow-guide.md: command selection, mode choice, and example flowsreferences/prompt-caching.md: stable-prefix ordering, cache telemetry, and cache-safe prompt structurereferences/evaluation.md: how to benchmark the skill and interpret results
Eval scaffolding
Starter prompts live in evals/evals.json. Use them when iterating on the skill or when you want a small repeatable benchmark set.
Integration with agent-studio
Compression trigger pipeline
compression-trigger.cjsdetects context pressure → writescompression-reminder.txt- Router reads reminder → spawns
context-compressoragent with this skill - Agent runs
run_skill_workflow.pywith the query context - Real compression stats logged to
compression-stats.jsonl
Memory persistence
After compression, persist distilled learnings via MemoryRecord:
gotchas.json: text contains gotcha|pitfall|anti-pattern|risk|warning|failureissues.md: text contains issue|bug|error|incident|defect|gapdecisions.md: text contains decision|tradeoff|choose|selected|rationalepatterns.json: default fallback for all remaining distilled evidence
ccusage cost tracking
Read .claude/context/runtime/ccusage-status.txt for live token usage before deciding compression aggressiveness. This file is auto-updated by ccusage-statusline.cjs on every tool use. Format:
[tokens] 135,345 today (in: 14,850 / out: 120,495) | Cost: $127.4826
[cache] $627.2992 saved | 139,399,832 reads, 8,751,364 writes
The Router MUST display this at every pipeline milestone (P0 user feedback). Fallback: ccusage --no-color 2>&1 | tail -5.
Iron Laws
- ALWAYS use the exact script commands above — never fall back to generic CLAUDE.md guidance
- ALWAYS profile first before compressing to understand the input size
- NEVER compress without a query when correctness matters — use
evidence_awaremode - ALWAYS persist distilled learnings via MemoryRecord after compression
- NEVER bluff if evidence is insufficient — recommend a broader pass
- ALWAYS report savings, risks, and next safest action in plain language
Related Skills
oimiragieo/neurokit2
tools
VerifiedTrustedCommunity
Comprehensive biosignal processing toolkit for analyzing physiological data including ECG, EEG, EDA, RSP, PPG, EMG, and EOG signals. Use this skill when processing cardiovascular signals, brain activity, electrodermal responses, respiratory patterns, muscle activity, or eye movements. Applicable for heart rate variability analysis, event-related potentials, complexity measures, autonomic nervous system assessment, psychophysiology research, and multi-modal physiological signal integration.
24SKILL.mdUpdated Apr 15, 2026
oimiragieo/networkx
tools
VerifiedTrustedCommunity
Comprehensive toolkit for creating, analyzing, and visualizing complex networks and graphs in Python. Use when working with network/graph data structures, analyzing relationships between entities, computing graph algorithms (shortest paths, centrality, clustering), detecting communities, generating synthetic networks, or visualizing network topologies. Applicable to social networks, biological networks, transportation systems, citation networks, and any domain involving pairwise relationships.
24SKILL.mdUpdated Apr 15, 2026
oimiragieo/molfeat
data-ai
VerifiedTrustedCommunity
Molecular featurization for ML (100+ featurizers). ECFP, MACCS, descriptors, pretrained models (ChemBERTa), convert SMILES to features, for QSAR and molecular ML.
24SKILL.mdUpdated Apr 15, 2026
oimiragieo/modal
development
VerifiedTrustedCommunity
Run Python code in the cloud with serverless containers, GPUs, and autoscaling. Use when deploying ML models, running batch processing jobs, scheduling compute-intensive tasks, or serving APIs that require GPU acceleration or dynamic scaling.
24SKILL.mdUpdated Apr 15, 2026