forge

Execute a defined research experiment with traceable inputs, configurations, runs, and outputs. Use for experiment implementation and reproducibility, not to invent the research claim.

Inputs

Versioned experimental plan; implementation; data access; environment; compute authorisation and budget.

Run record

Use checklist 0.5 or newer for this SOP. Reading or reloading the skill must not reset work. For a new task, run checklist init "${CLAUDE_SKILL_DIR}" --new --path "/absolute/project"; retain its returned ID as RUN. For existing work, use checklist resume "$RUN", then checklist show --run "$RUN". Outside Claude, replace the skill-directory variable with this directory’s actual path.

Commands below record work after it is done. Replace evidence placeholders with concise artifact locations and observations. Manual confirmations are not independent proof; a sensor pass covers only the command and inputs it observed. verify does not close a stage; advance checks that its required readings are fulfilled. Use report --run "$RUN" for the retained history.

Procedure

prepare

Match code, data, and configuration to the plan. Run a small end-to-end sanity check and verify that metrics and failure records have the intended meaning.

Record these artifacts after doing the work:

• Code, data, environment, and plan versions are recorded. checklist check prepare inputs --run "$RUN" --evidence "<artifact and observation>"
• A small end-to-end run validates data flow and metric computation. checklist check prepare sanity --run "$RUN" --evidence "<artifact and observation>"

When this stage is fulfilled: checklist advance prepare --run "$RUN".

execute

Run within the agreed budget. Preserve seeds, commands, logs, failed runs, and deviations; checkpoint where useful. Keep exploratory adjustments separate from the committed comparison.

Record these artifacts after doing the work:

• Successful and failed runs have reproducible configurations and logs. checklist check execute runs --run "$RUN" --evidence "<artifact and observation>"
• Plan deviations and data-boundary checks are recorded. checklist check execute integrity --run "$RUN" --evidence "<artifact and observation>"

When this stage is fulfilled: checklist advance execute --run "$RUN".

package

Regenerate at least a representative result from saved inputs and document the remaining reproduction cost. Package raw-to-summary transformations, not only final charts.

Record these artifacts after doing the work:

• A representative result was regenerated from recorded inputs. checklist check package reproduction --run "$RUN" --evidence "<artifact and observation>"
• Run manifest and analysis inputs are available for independent review. checklist check package artifacts --run "$RUN" --evidence "<artifact and observation>"

When this stage is fulfilled: checklist advance package --run "$RUN".

Branches

If resources or data are unavailable, report the blocked run rather than substitute synthetic results silently. A method change that changes the comparison returns to ledger for a new plan version.

If an input or permission is missing, name the blocker and leave the relevant item pending. A changed requirement may need a new run; an accepted checklist edit uses resume "$RUN" --refresh and requires fresh readings. To abandon work, use checklist reset --run "$RUN" --reason "<reason>"; history is retained.

Completion

A traceable experiment artifact set for reckoning and envoy; resource use stays within explicit authorisation.

After all required work and stage closures is recorded, use checklist done --run "$RUN". This archives the result without deleting its history; it does not authorise any external action.

Reference shelf

Open the technique relevant to the current step. These notes provide concrete methods and examples, not additional phases; this entrypoint defines the run's completion conditions.

• Preserve the identity and failures of a research run
• Validate the path from input data to reported metric
• Regenerate a result from retained artifacts
• Execute a plan without hiding methodological changes