feat: add CLI history baseline and regression ergonomics

README.md

@@ -108,6 +108,17 @@ Write a timestamped local report bundle:
 bun run --filter '@ignitionai/agent-trainer-cli' dev -- eval run ./examples/context-engineering/experiment.ts --bundle reports
 ```
 
+Record local experiment history and run a regression check against the latest baseline:
+
+```bash
+bun run --filter '@ignitionai/agent-trainer-cli' dev -- eval run ./examples/context-engineering/experiment.ts \
+  --history .ignition/experiment-history.jsonl \
+  --baseline latest \
+  --regression \
+  --max-score-drop 0.03 \
+  --record-history
+```
+
 Run the sample CI regression gate:
 
 ```bash

docs/ALPHA_READINESS.md

@@ -27,7 +27,7 @@ All packages declare `license: MIT`, matching the root `LICENSE` file.
 | `@ignitionai/agent-trainer-adapter-langgraph` | ready | ready | ready | ready | partial | Structural adapter only; no persistence, streaming or graph internals. |
 | `@ignitionai/agent-trainer-adapter-mastra` | ready | ready | ready | ready | partial | Structural adapter only; no memory, tool or full Mastra coverage. |
 | `@ignitionai/agent-trainer-adapter-vercel-ai` | ready | ready | ready | ready | partial | Structural adapter only; no streaming, tools or live provider calls. |
-| `@ignitionai/agent-trainer-cli` | ready | ready | ready | ready | partial | Runs typed experiments and writes standalone reports or timestamped bundles; no history/baseline/regression flags yet. |
+| `@ignitionai/agent-trainer-cli` | ready | ready | ready | ready | partial | Runs typed experiments, writes reports/bundles, records local history, selects baselines and runs regression checks; no watch mode or remote execution. |
 | `@ignitionai/agent-trainer-core` | ready | ready | ready | ready | partial | Foundational helpers have dedicated tests; runtime schema validation remains outside the current helper surface. |
 | `@ignitionai/agent-trainer-environment` | ready | ready | ready | ready | partial | Tested episode runner with safety guards and a deterministic RAG episode example; no production runtime or optimization loop. |
 | `@ignitionai/agent-trainer-evals` | ready | ready | ready | ready | partial | Current rewards are tested; RAG presets and richer scoring are still missing. |
@@ -55,7 +55,6 @@ All packages declare `license: MIT`, matching the root `LICENSE` file.
 
 ## Known Work After Internal Alpha
 
-- Add CLI history, baseline selection and regression command ergonomics after report bundles and CI examples.
 - Add deeper mocked examples for ecosystem adapters.
 - Dogfood the alpha inside IgnitionRAG and collect trajectory/reward evidence.
 - Add a lightweight policy optimization loop after real dogfood produces trajectory data.

docs/BACKLOG.md

@@ -2793,7 +2793,7 @@ Next PR:
 
 Status:
 
-- current
+- completed
 
 Branch:
 
@@ -2850,7 +2850,7 @@ Next PR:
 
 Status:
 
-- planned
+- current
 
 Branch:
 

docs/PROJECT_AUDIT.md

@@ -75,12 +75,12 @@ If a package exists but is intentionally narrow, minimal or untested, it is part
 
 ### `@ignitionai/agent-trainer-cli`
 
-- Purpose: run typed experiment modules locally and write JSON/Markdown reports or report bundles.
+- Purpose: run typed experiment modules locally, write JSON/Markdown reports or report bundles, record local history and run baseline regression checks.
 - Main exports: `parseCliArgs`, `runCli`, `CliCommand`, `CliEnvironment`.
 - Stability level: stable for the current local CLI surface.
 - Tests present: yes.
 - Example present: yes, `examples/context-engineering/experiment.ts` through the CLI.
-- Known limitations: no watch mode, no persistent history flag, no baseline selection flag, no regression-gate command, no remote execution.
+- Known limitations: no watch mode, no hosted history, no remote execution and no provider-backed regression scoring.
 
 ### `@ignitionai/agent-trainer-core`
 

packages/cli/README.md

@@ -60,6 +60,60 @@ reports/
    └─ metadata.json
 ```
 
+## Keep Local History
+
+Use `--history` with `--record-history` to append the result to a local JSONL history file:
+
+```bash
+bun run --filter '@ignitionai/agent-trainer-cli' dev -- eval run ./examples/context-engineering/experiment.ts \
+  --history .ignition/experiment-history.jsonl \
+  --record-history
+```
+
+List recent entries:
+
+```bash
+bun run --filter '@ignitionai/agent-trainer-cli' dev -- eval history list .ignition/experiment-history.jsonl \
+  --experiment context-engineering-strategies \
+  --limit 5
+```
+
+Inspect the latest entry for an experiment:
+
+```bash
+bun run --filter '@ignitionai/agent-trainer-cli' dev -- eval history show .ignition/experiment-history.jsonl latest \
+  --experiment context-engineering-strategies
+```
+
+The history file is newline-delimited JSON using `ignition.experiment-history-entry.v1`.
+
+## Run Regression Checks
+
+Use `--baseline latest` to compare the current run against the latest matching history entry.
+`--regression` makes the command fail when the comparison exceeds the allowed thresholds:
+
+```bash
+bun run --filter '@ignitionai/agent-trainer-cli' dev -- eval run ./examples/context-engineering/experiment.ts \
+  --history .ignition/experiment-history.jsonl \
+  --baseline latest \
+  --regression \
+  --max-score-drop 0.03 \
+  --max-latency-increase-ms 250 \
+  --max-cost-increase-usd 0.002 \
+  --regression-markdown reports/regression.md
+```
+
+You can also pass a concrete history entry id instead of `latest`:
+
+```bash
+bun run --filter '@ignitionai/agent-trainer-cli' dev -- eval run ./examples/context-engineering/experiment.ts \
+  --history .ignition/experiment-history.jsonl \
+  --baseline context-engineering-strategies-2026-01-01T00-00-00-000Z \
+  --regression
+```
+
+Use `--variant <id>` one or more times when only specific variants should be checked.
+
 ## Non-goals
 
-The CLI does not implement watch mode, remote execution, hosted dashboards, auth, provider keys, persistent history or regression gates.
+The CLI does not implement watch mode, remote execution, hosted dashboards, auth, provider keys, hosted history or provider-backed regression scoring.