Retrieval, structured extraction, tools, memory, and media boundaries. Start with the decision you need to make, then use the linked material to build a checkable result.
Learning path
- Retrieved context: select sufficient evidence before writing the answer
- Evidence IDs: make generated claims auditable against supplied records
- Long context: make inclusion and truncation testable
- Document extraction: separate observed fields from inferred values
- Tool calls: validate intent and arguments before an external effect
- Tool loops: set budgets, state checks, and a stopping condition
- Conversation memory: retain decisions without retaining every private detail
- Multimodal prompts: separate what an image shows from what it suggests
Connect the path
The foundations define the task and trust boundary. Later modules test whether a result should be used, revised, or sent to a person.
- Prompt Engineering
- Production prompt engineering
- Prompt engineering projects
- Prompt engineering quizzes
Prompt construction and response boundaries
These lessons turn prompt wording into a tested workflow. Apply each rule to a real input boundary, then inspect whether the decision and any side effect still match the task contract.
- Prompt templates: bind variables without changing instruction structure
- Instruction conflicts: resolve authority before wording
- Model settings: change one generation variable against a fixed case set
- Streaming responses: validate the complete artifact before an effect
- Retrieved evidence: reconcile versions and conflicting facts
Queries, complete evidence, and exact computation
A reliable answer needs the right information, a valid calculation or decision, and a destination that treats generated content as data.
- Retrieval queries: resolve references without adding facts
- Document chunks: preserve the clause and its governing exception
- Numeric prompts: let code calculate and the model explain
Agent workflow boundaries
Select tools, schedule dependent calls, repair structure, and resume from verified state.
- Tool catalogs: describe eligibility, inputs, and effects
- Tool plans: separate independent reads from dependent actions
- Structured outputs: repair format without changing the decision
- Conversation checkpoints: resume from verified state
- Agent workflow decisions
- Project: recover a tool workflow without duplicate effects
Reasoning with explicit checks
Use visible grounds, candidate filters, bounded critique, boundary examples, and source-preserving context.
- Reasoning summaries: show checkable grounds, not invented certainty
- Multiple candidates: filter invalid answers before ranking
- Critique loops: require a named defect and a stopping rule
- Few-shot examples: teach the boundary with near misses
- Context distillation: shorten input without losing governing exceptions
- Project: verify a contract-renewal prompt at the boundary
- Reasoning and evidence checks
Bounded reasoning methods
Compare sampling, branch search, observation loops, ordered decomposition, and executable calculations.
- Self-consistency: sample answers, then verify the winner
- Tree of Thoughts: branch only where a decision can be checked
- ReAct: alternate tool actions with checked observations
- Least-to-most prompting: solve smaller dependencies first
- Program-aided prompting: make computation executable and bounded
- Project: select and verify a reasoning pattern
- Reasoning patterns and operating limits