Run the first AletheIA test

Validate the macro-governance and micro-execution split with one small, evidence-producing task.

This exercise tests whether AletheIA and Adaptive Skills remain understandable when used together. It is an operating test, not a benchmark or runtime integration.

Before you begin

Choose one small feature-like task that:

• has a clear human owner;
• can be completed in one bounded Work Slice;
• produces visible proof;
• does not require sensitive or restricted context for the first trial.

Use this initial skill set:

• workflow;
• feature-planning;
• testing.

Add triad-check only when the task genuinely crosses product, design, and engineering.

Step 1 — Let AletheIA frame the task

Record:

• the goal;
• the immediate boundary;
• explicit non-goals;
• minimum closure evidence;
• review or handoff gates;
• known uncertainty.

Do not select a large skill bundle during framing.

Step 2 — Use skills for micro execution

Run one skill at a time:

1. workflow makes the local execution boundary and next step explicit.
2. feature-planning defines the smallest useful delivery slice.
3. testing defines proof proportionate to risk.

Inspect each output before invoking the next skill. A skill may expose a reason to stop, review, or change the plan.

Step 3 — Return evidence

Capture:

• which skill and modules were used;
• the usable, partial, failed, or unavailable result;
• authoritative evidence references;
• known risks and missing values;
• the advisory handoff signal.

Use the skill observation return pattern when a structured return helps. Do not infer unavailable values.

Step 4 — Let AletheIA evaluate closure

The governing layer checks:

• whether the selected skills matched the Work Slice;
• whether evidence satisfies the declared gate;
• whether a human review or handoff is required;
• whether a repeated failure should become an observation;
• whether the slice should continue, change, escalate, or close.

The skill output informs this decision but does not make it.

Record what you learned

At the end, write:

• what AletheIA governed;
• what each skill contributed;
• what remained consumer-local;
• what felt clear or redundant;
• what evidence was available or missing;
• the next safe improvement, if any.

Use the repository’s first-test report template in your local copy: examples/aletheia/first-test-report.md.

Success criteria

The trial succeeds when:

• the macro/micro split remains legible;
• framing and proof improve with low overhead;
• the skill set stays small;
• evidence remains source-backed;
• project-local rules are not mistaken for generic library behavior.

Failure signals

Stop and review when:

• AletheIA duplicates skill internals;
• a skill starts deciding macro gates or closure;
• proof ownership becomes unclear;
• the setup is disproportionate to the task;
• the workflow requires invented evidence or unavailable measurements.

Next steps

• Compare the result with Adaptive Skills and AletheIA.
• Read Observation and evidence return before formalizing the record.
• Use Workflow recipes for later tasks, while preserving AletheIA as the governing frame.