.agents/skills/auto-iterate-goal/SKILL.md
Generate or validate the auto-iterate goal file before launching WF8 auto-iterate
npx skillsauth add linzhe001/Harness-Research auto-iterate-goalInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
Bridge skill between WF7.5 PASS and WF8 auto-iterate start. Produces or validates docs/auto_iterate_goal.md — the operator-facing research objective that the controller's goal parser consumes.
$validate-run returns PASS (WF7.5 gate cleared)tooling/auto_iterate/scripts/auto_iterate_ctl.sh start --goal docs/auto_iterate_goal.md./references/goal-template.md.agents/references/workflow-guide.md../evaluate/references/stage-report.mdtooling/auto_iterate/docs/auto_iterate_goal_template.mdinitGenerate docs/auto_iterate_goal.md when it does not exist.
Sources:
CLAUDE.md / PROJECT_STATE.jsonOutput: docs/auto_iterate_goal.md with all required structured fields.
refreshUpdate an existing goal when fields are missing, defaults changed, or new context is available.
Behavior:
docs/auto_iterate_goal.mdcheckValidate the existing goal file without modifying it.
Checks:
primary_metric.direction is maximize or minimizebudget.max_rounds is a positive integerscreening_policy.enabled is boolean{{...}} markers remainOutput: PASS or list of validation errors.
.auto_iterate/goal.md (that is the controller's job via start --goal).auto_iterate/state.jsonNEXT_ROUND / CONTINUE / etc. (that is $evaluate's job)The orchestrator should auto-trigger after WF7.5 PASS:
WF7.5 validate-run PASS
→ $auto-iterate-goal check
→ goal exists + valid: no-op
→ goal missing: $auto-iterate-goal init
→ goal invalid/incomplete: $auto-iterate-goal refresh
development
WF7.5 training pipeline validation. Before entering WF8 iteration, first use Codex to review code for baseline equivalence, then run a 100-step smoke test to verify end-to-end pipeline functionality.
business
WF1 Inspiration survey and gap analysis. Takes the user's research idea, performs literature search, gap analysis, competitor analysis, and feasibility scoring, then outputs Feasibility_Report.md. Use when the user has a new CV research idea that needs a feasibility assessment.
tools
WF10 Submission/Release Tool. Multi-scene training, result packaging, filename validation, dry-run submission checks. Used after ablation experiments are complete and before competition submission.
development
WF2 Architecture refinement and MVP design. Reads the feasibility report, analyzes the base codebase architecture, designs plug-and-play new modules, defines the MVP, provides A/B/C alternative plans, and outputs Technical_Spec.md. Use when a research idea needs to be translated into a concrete technical architecture design.