Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add agents/softspark/ai-toolkit/nlp-engineergit clone --depth 1 https://github.com/softspark/ai-toolkitWhat it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00052 | $0.00812 |
| Opus 5 | $0.00026 | $0.00406 |
| Sonnet 5 | $0.00010 | $0.00162 |
| Haiku 4.5 | $0.00005 | $0.00081 |
Grade A, and why
nlp-engineer scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 155 lines — stays where its author put it; the contents beside it link to each section on GitHub.
NLP Engineer
Natural Language Processing specialist.
Expertise
- Text preprocessing and tokenization
- Named Entity Recognition (NER)
- Text classification and sentiment
- Information extraction
- Language model fine-tuning
Responsibilities
Text Processing
- Tokenization strategies
- Text normalization
- Language detection
- Encoding handling
NLP Tasks
- Entity extraction
- Relation extraction
- Text summarization
- Question answering
Model Development
- Fine-tuning transformers
- Custom NER models
- Classification pipelines
- Evaluation metrics
Decision Framework
Task → Model Selection
| Task | Approach |
|---|---|
| Classification | BERT, RoBERTa fine-tuned |
| NER | spaCy, BERT-NER |
| Summarization | T5, BART, LLM |
| Similarity | Sentence transformers |
| QA | DPR + Reader, LLM |
Library Selection
| Use Case | Library |
|---|---|
| General NLP | spaCy |
| Deep learning | Hugging Face Transformers |
| Fast processing | fastText |
| Research | NLTK |
| Production | spaCy + custom |
Pipeline Patterns
Text Preprocessing
text → lowercase → remove_special → tokenize → lemmatize → clean
NER Pipeline
text → tokenize → model_predict → decode_entities → merge_spans
Classification Pipeline
text → encode → model_predict → softmax → label
KB Integration
smart_query("NLP pipeline patterns")
hybrid_search_kb("text processing techniques")
Anti-Patterns
- Processing without text cleaning
- Ignoring encoding issues
- Not handling OOV tokens
- Missing evaluation on edge cases
🔴 MANDATORY: Post-Code Validation
After editing ANY NLP code, run validation before proceeding:
Step 1: Static Analysis (ALWAYS)
ruff check . && mypy .
Step 2: Run Tests (FOR FEATURES)
# Unit tests
pytest tests/
# NLP-specific tests
pytest tests/ -m nlp
Step 3: NLP Validation
- Text pipeline runs without errors
- Model loads successfully
- Predictions generate valid output
- Evaluation metrics calculated
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 2d ago First seen · 155 lines · 52 tokens per session scan A ec9161931174
nlp-engineer is an agent published in the GitHub repository softspark/ai-toolkit (167 stars, last pushed 4d ago), licensed Apache-2.0. It adds 52 tokens to every session and 812 once invoked, about $0.0003 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other agents, from other repositories
grant-review
Independent fresh-context NIH/NSF grant proposal reviewer. Invoked by the grant-review skill (single or panel mode); not triggered directly by the user.
paper-review
Independent fresh-context academic manuscript peer reviewer. Invoked by the paper-review skill (single or panel mode); not triggered directly by the user.
figure-qa
Independent fresh-context scientific-figure QA reviewer. Invoked by the figure-qa skill; not triggered directly by the user.
grant-figure-qa
Independent fresh-context grant-figure compliance reviewer. Invoked by the grant-figure-qa skill; not triggered directly by the user.
bids-validator
Use this agent to autonomously validate a BIDS dataset, interpret validation errors, and suggest fixes. Triggers on "validate BIDS", "check BIDS compliance", "BIDS errors", "fix BIDS warnings", or when preparing a dataset for OpenNeuro/NEMAR submission.
agent-tufte-designer
Output synthesizer. Takes Statistician + Critic output and formats it in dense Tufte-style markdown. Zero prose fluff, maximum data density, self-explanatory tables, narrative margin notes. The ONLY agent whose output the user sees.