plan

<objective> Wizard: scans codebase, proposes metric/guard/agent config, writes `program.md` run spec. Also runs cProfile on file path to surface bottlenecks before prompting for optimization goal. NOT for: running experiments (use `/research:run`); methodology validation (use `/research:judge`); full pipeline from goal to result (use `/research:sweep`); benchmarking or microbenchmark design (use `foundry:perf-optimizer` — plan's scope is ML metric optimization loops, not raw latency/throughput benchmarking). </objective> <workflow> <!-- Agent resolution: see _RESEARCH_SHARED/agent-resolution.md --> ## Agent Resolution **Environment precondition** — `CLAUDE_PLUGIN_ROOT` set automatically when skill runs via plugin manager. Fallback `plugins/cc_research` resolves only from project root. Neither resolves (bare `.claude/` copy invoked from subdirectory) → bin/ scripts return empty strings silently. **bin/ scripts this skill depends on** (deployed inside `${CLAUDE_PLUGIN_ROOT}/bin/`): `resolve_shared.py`, `make_run_dir.py`. Each call below followed by explicit empty-result guard — silent failure surfaces as fail-fast error, never empty-string path. **Agent resolution**: load and follow the protocol below. Contains: foundry check + fallback table. Foundry not installed → substitute each `foundry:X` with `general-purpose` per table. Agents this skill uses: `foundry:solution-architect`, `foundry:perf-optimizer`, `research:scientist`. ```bash export CSID="${CLAUDE_CODE_SESSI...