Package index • LLMRagent

Agents, budgets, and memory

Create agents with model configurations, personas, memory policies, and declared budgets.

agent(): Create an agent
budget(): Spending and effort limits for an agent
memory_buffer() memory_summary() memory_recall(): Agent memory policies

Declare tool effects and limits, check boundaries, and pause tool calls for a human decision.

agent_tool(): Define a governed tool
guardrail(): Define a guardrail
guardrails(): Collect guardrails
human_gate(): Mark a point or a tool as requiring human approval
approve_tool_call(): Approve, reject, or edit a pending tool call
resume_run() as.character(<agent_resume_result>) print(<agent_resume_result>): Resume a paused run after a tool-approval decision

Run general conversations or formats designed for argument, group discussion, interviews, and collective decisions.

Run factorial designs and robustness batteries, and check cross-cell state leakage.

agent_experiment() print(<agent_experiment>): Run a factorial agent experiment
check_state_leakage(): Detect shared state across experiment cells
agent_robustness(): Run a robustness battery
vary_models() vary_temperature() vary_prompt() vary_persona() vary_option_order(): Robustness perturbation axes

Represent and vary personas, inspect essentializing language, and mark the scope of claims.

persona_frame() print(<persona_frame>) as.character(<persona_frame>): A persona as an auditable research object
persona_variants() print(<persona_set>): Vary a persona along named dimensions
persona_audit() print(<persona_audit>): Audit persona briefs for essentializing language and caricature
mark_claim_type(): Mark the kind of claim a run can support
llm_claim_lint(): Assert (or scope) prose against a run's claim type

Coordinate specialist agents through model-directed delegation, fixed pipelines, or fan-out synthesis.

agent_as_tool(): Expose an agent as a tool for other agents
agent_pipeline(): Run input through a chain of agents
agent_fanout_synthesis(): Work a hard problem with one strong model and many cheap ones

Create common run records, manifests, and inspectable archives. Includes report, diagnostics, and reset methods for LLMRagent results.

as_agent_run() as_tibble(<agent_run>): Convert an LLMRagent result to a unified run object
agent_manifest(): Build the study manifest for a run
archive_agent_study(): Seal an agent study to a directory
hash_persona(): Hash a persona
hash_tool_spec(): Hash a tool's declared specification
llmragent-methods: LLMR-family methods for LLMRagent run objects
reexports diagnostics report reset: Objects exported from other packages
diagnostics(<agent_run>) diagnostics(<Agent>): Machine-readable diagnostics for an agent run
diagnostics(<agent_experiment>): Machine-readable diagnostics for an agent experiment
diagnostics(<persona_audit>): Machine-readable diagnostics for a persona audit
report(<agent_run>) report(<Agent>): Draft a methods-section report for an agent run
report(<agent_experiment>): Draft a short report for an agent experiment
reset(<Agent>): Clear an agent's memory

Branched and resumable procedures with checkpoints, forks, and workflow replay.

Connect governed MCP tools and write an HTML view of a run record.

Persist agent state and restore it with tool functions attached again.

LLMRagent LLMRagent-package: LLMRagent: agents, multi-agent conversations, and agent experiments