run— agent skill

<objective>

Sustained metric-improvement loop — reads program.md, iterates specialist ideation agents, commits atomically, auto-rolls back on regression. For long-running automated improvement campaigns.

NOT for: methodology validation before run (use /research:judge); hypothesis generation (use research:scientist agent); one-off feature work (use /develop:feature).

</objective>

<constants>

Campaign mode only:

MAX_ITERATIONS:             20 (hard cap — program.md values above 20 are silently clamped with a warning; 20 is the effective maximum)
MAX_CODEX_RUNS:             10 (cost ceiling for --codex Phase 2c — disable Codex once exceeded)
STUCK_THRESHOLD:            5 consecutive discards → escalation
GUARD_REWORK_MAX:           2 attempts before revert
VERIFY_TIMEOUT_SEC:         120 (local), 300 (--colab)
COLAB_KNOWN_HW:             H100, L4, T4, A100
SUMMARY_INTERVAL:           10 iterations
DIMINISHING_RETURNS_WINDOW: 5 iterations < 0.5% each → warn user and suggest stopping
STATE_DIR:                  .experiments/state/<run-id>/  (timestamped dir per run — see .claude/rules/artifact-lifecycle.md)
SENTINEL_SLUG_FORMULA: |
  eval "$(bash "${CLAUDE_PLUGIN_ROOT:-plugins/research}/bin/git_slugs.sh")"
  # Sentinel path: ${TMPDIR:-/tmp}/claude-commit-auth-${REPO_SLUG}-${BRANCH_SLUG}
  # Bash state is lost between tool calls — re-source git_slugs.sh at each use site; it is the only authorized slug form.

<!-- Note: STATE_DIR (.experiments/state/) holds per-iteration artifacts (diary, experiments.jsonl). Hypothesis pipeline outputs (hypotheses.jsonl, checkpoint.json, journal.md) go to .experiments/<run-id>/ (RUN_DIR). These are two separate directories by design — see protocol.md for layout. -->

Agent strategy mapping (agent_strategy in config → ideation agent to spawn):

Auto-inference keyword heuristics (when agent_strategy: auto or omitted; checked against ## Goal text AND metric command):

Precedence order (first match wins; ML keywords take priority over test-framework keywords). ML-specific compound terms (not bare tokens) required to prevent over-triggering on eval/train/val as common words:

• contains accuracy, loss (when paired with train_loss/val_loss/eval_loss), f1_score, auc_roc, auroc, train_step, val_acc, eval_loss, epoch, gradient, tensor, overfit, generaliz, regulariz, validation, dropout, weight_decay, lr_schedule, cross_val, precision, recall, OR explicit --scientist flag → ml → research:scientist
• contains time, latency, bench, throughput, memory → perf → foundry:perf-optimizer
• contains pytest, coverage, complexity → code → foundry:sw-engineer
• no keyword match → perf (default fallback) — WARN: print ⚠ No keyword match — defaulting to 'perf' strategy. If this is an ML task, set agent_strategy: ml in program.md. Log resolved agent + reason in state.json strategy_resolution.

Bare tokens eval, train, val (without compound suffix) do NOT trigger ml routing — too common in non-ML contexts (test eval scripts, training-environment configs, validator command names).

Stuck escalation sequence (at STUCK_THRESHOLD consecutive discards):

1. Switch to different agent type. Rotation by current strategy:

1. Spawn 2 agents parallel with competing strategies; each writes full analysis to .experiments/state/<run-id>/stuck-escalation-<i>-<agent-type>.md, returns ONLY compact JSON envelope. Use this spawn prompt verbatim (substitute <run-id>, <i>, and strategy):

   Stuck-escalation handoff — iteration <i> after STUCK_THRESHOLD consecutive discards.
   Read `.experiments/state/<run-id>/state.json` for goal, best_metric, baseline, config.
   Read `.experiments/state/<run-id>/experiments.jsonl` for full iteration history.
   Read `.experiments/state/<run-id>/diary.md` for qualitative context (what was tried, why reverted).
   Read `.experiments/state/<run-id>/context-<i>.md` for current iteration's context block.
   Continue from the last completed iteration (do NOT restart from iteration 0).
   Write your full analysis and proposed change to `.experiments/state/<run-id>/stuck-escalation-<i>-<your-strategy>.md`.
   Write a resume point to `.experiments/state/<run-id>/resume.json`: {iteration: <i>, strategy: "<your-strategy>", proposed_change: "<one-line description>"}.
   Return ONLY: {"strategy":"<your-strategy>","description":"...","files_modified":[...],"confidence":0.N,"file":".experiments/state/<run-id>/stuck-escalation-<i>-<your-strategy>.md"}

Consolidation: pick whichever returns delta ≥ 0.1% AND guard pass; if both qualify, pick higher delta.

1. Stop, report progress, surface to user — no blind looping

</constants>

<workflow>

<!-- Agent resolution: see _RESEARCH_SHARED/agent-resolution.md -->

Agent Resolution

_RESEARCH_SHARED=$(python "${CLAUDE_PLUGIN_ROOT:-plugins/research}/bin/resolve_shared.py" 2>/dev/null)  # timeout: 5000
[ -z "$_RESEARCH_SHARED" ] && { echo "! Plugin path resolution failed — ensure research plugin installed and CLAUDE_PLUGIN_ROOT set, or invoke from project root."; exit 1; }

`CLAUDE_SKILL_DIR` resolution — constants block provides default plugins/research/skills/run (source-tree path). Resolve to installed path before use:

CLAUDE_SKILL_DIR=$(ls -td ~/.claude/plugins/cache/borda-ai-rig/research/*/skills/run 2>/dev/null | head -1)
[ -z "$CLAUDE_SKILL_DIR" ] && CLAUDE_SKILL_DIR="$(git rev-parse --show-toplevel 2>/dev/null)/plugins/research/skills/run"

Read $_RESEARCH_SHARED/agent-resolution.md. Contains: foundry check + fallback table. If foundry not installed: use table to substitute each foundry:X with general-purpose. Agents this skill uses: foundry:sw-engineer, foundry:linting-expert, foundry:perf-optimizer, foundry:solution-architect, research:scientist.

Default Mode (Steps R1–R7)

Triggered by run <goal|file.md>.

Task tracking: create tasks R0–R7 at start. If no --researcher/--architect, mark R0 skipped. If --codex active, create task R5b: Codex co-pilot (iter ?/max) status pending.

Step R0: Hypothesis pre-phase (--researcher / --architect)

If no --researcher/--architect, skip to R1.

Flag combination guard: if --researcher is set but --architect is NOT, print ⚠ --researcher without --architect: hypotheses will NOT be validated for architectural feasibility before execution — infeasible hypotheses may waste iterations. Add --architect for feasibility filtering. then continue (this is advisory, not blocking). If only --architect is set without --researcher, the feasibility filter applies to oracle-generated hypotheses — valid combination.

Read ${CLAUDE_SKILL_DIR}/modes/hypothesis-pipeline.md

Per-iteration hypothesis selection (when --researcher/--architect set, inside R5 loop): pop next from RESEARCH_QUEUE. Append to Phase 2 prompt: "Focus this iteration on testing this hypothesis: <hypothesis text>."

Per-iteration journal hook (inside R5, after Phase 7): if --journal active, append entry to <RUN_DIR>/journal.md after EVERY iteration — regardless of outcome. Entry format: protocol.md (companion file, same skill dir). # loads: protocol.md Journals record kept and reverted iterations so ideation agent learns failed approaches.

Per-iteration checkpoint write (after Phase 7): if --researcher/--architect active, append one line to <RUN_DIR>/checkpoint.json per schema in protocol.md (companion file, same skill dir): {iteration, hypothesis_id, metric_before, metric_after, status: "passed"|"rolled_back"}.

Step R1: Load / build config

`--resume` flag detection: if --resume in args, extract optional program.md path. Jump to ## Resume Mode. Rest of R1 and R2–R7 skipped.

`--hypothesis <path>` parsing: if --hypothesis in args, extract path token following it. Verify file exists: [ -f "$HYPOTHESIS_PATH" ]. If not found: print ! --hypothesis <path>: file not found and stop. If found: set hypothesis_override = true. In R5 Phase 2 (Propose change), replace the oracle-generated hypothesis with the loaded file content — prepend to the ideation agent prompt: "Use this pre-specified hypothesis as your starting hypothesis for iteration N: <contents of HYPOTHESIS_PATH>. Validate, refine, and implement it. Do not generate a new hypothesis from scratch."

Auto-detect: first non-flag arg ends in .md → parse as program file. Otherwise → text goal.

Clarification prompt (.md file only): after extracting .md path, inspect next token (before -- flags):

• If absent or starts with -- → clarification_prompt = null
• Quoted string (starts and ends with ") → extract as clarification_prompt, strip quotes
• Bare unquoted token (no --, no ") → accept as clarification_prompt; print: ℹ clarification set to "<token>" (tip: quote multi-word hints — e.g. "/research:run program.md \"focus on sort\" --codex")

After clarification extraction, remaining non-flag tokens (not starting --) are unrecognized. For each, print:

⚠ Unrecognized argument "<token>" — ignored.
  Known positional args: <program.md path> [clarification]
  Known flags: --team, --colab[=HW], --codex, --compute=local|colab|docker, --researcher, --architect, --journal, --hypothesis <path>
  If you meant to override the algo, edit the ## Config block in your program.md (algo: sort) and update ## Metric to match.
  If you meant to set a clarification hint, pass it as a quoted string: "/research:run program.md \"sort improvements\" --codex"

Unsupported flag check: follow $_RESEARCH_SHARED/unsupported-flag-protocol.md. Supported flags for this skill: --resume, --team, --compute, --colab, --codex, --researcher, --architect, --journal, --hypothesis, --scientist.

If argument is a `.md` file — read and parse with these rules:

1. Find each ## <Section> heading (case-insensitive).
2. Extract first fenced code block following that heading.
3. Parse block as key: value lines; multi-value = indented - value items. Paths with spaces: wrap in double quotes.
4. Missing required fields (command under ## Metric/## Guard) → stop with error.
5. agent_strategy: auto (or omitted) → apply keyword heuristics from <constants> to ## Goal text and metric command.
6. target under ## Metric: direction: higher → stop when metric ≥ target; direction: lower → stop when metric ≤ target. If target omitted, run until max_iterations.
7. Unrecognized keys/headings → warn once, ignore.
8. ## Notes and # Program: title never parsed — human-only. (# Campaign: accepted as alias.)

If argument is text — auto-detect metric_cmd/guard_cmd from goal string and codebase scan (same as P-P1, non-interactive). config.json not read.

`--colab[=HW]` parsing: --colab (no =) → compute = "colab", colab_hw = null. --colab=<value> → compute = "colab", colab_hw = <value> (uppercased). Unknown <value> (not in {H100, L4, T4, A100}) → print "⚠ Unknown Colab hardware '<value>' — proceeding with default GPU. Known: H100, L4, T4, A100", set colab_hw = null. --compute=colab (no HW) → compute = "colab", colab_hw = null.

colab_hw in ## Config sets hardware preference (H100, L4, T4, A100); CLI --colab=HW overrides.

Generate run-id = $(date -u +%Y-%m-%dT%H-%M-%SZ). Assign immediately:

RUN_ID=$(date -u +%Y-%m-%dT%H-%M-%SZ)
RUN_DIR=".experiments/${RUN_ID}"  # hypothesis pipeline + journal outputs (per <constants> note)
STATE_DIR=".experiments/state/${RUN_ID}"  # per-iteration artifacts (state.json, experiments.jsonl, diary.md)
mkdir -p "$RUN_DIR" "$STATE_DIR"  # timeout: 5000 — both dirs created before any Write to either

Note: STATE_DIR (.experiments/state/${RUN_ID}/) is per-iteration artifact dir — distinct from RUN_DIR. Both coexist; see <constants> block.

Create run directory:

.experiments/state/<run-id>/
  state.json         ← iteration count, best metric, status
  experiments.jsonl  ← one line per iteration
  diary.md           ← human-readable research diary (hypothesis → outcome → decision)

Convert program_file to absolute path: realpath "$PROGRAM_FILE" — Resume Mode matches on absolute path.

Write initial state.json (program_file = absolute path to .md or null for text goal):

{
  "run_id": "<run-id>",
  "goal": "<goal>",
  "config": {},
  "program_file": "<absolute path to program.md, or null>",
  "iteration": 0,
  "best_metric": null,
  "best_commit": null,
  "status": "initializing",
  "started_at": "<ISO timestamp>",
  "clarification_prompt": null,
  "colab_hw": null,
  "sandbox_mode": "local"
}

Note: status is "initializing" until all R2 precondition checks pass — resume treats "initializing" as failed-init, not active run. Update to "running" at end of R2 (after all checks pass).

Step R2: Precondition checks

Run all checks before touching code. Fail fast with clear message:

1. Clean git: git status --porcelain → must be empty. If dirty: print dirty files and stop.
2. Not detached HEAD: git rev-parse --abbrev-ref HEAD → must not be HEAD.
3. Metric command numeric: run metric_cmd once; parse stdout for float. If no float: show output and stop.
4. Guard passes: run guard_cmd once; must exit 0. If fails: show output and stop.
5. `--colab` check: verify mcp__colab-mcp__runtime_execute_code available. If not, print setup instructions (see Colab MCP section) and stop. If --colab=HW (colab_hw non-null): print: Hardware requested: --colab=<colab_hw>. Ensure your Colab notebook running with <colab_hw> GPU.
6. `--codex` check: verify claude plugin list 2>/dev/null | grep -q 'codex@openai-codex'. If unavailable: print ⚠ codex plugin not found. Install it with: /plugin marketplace add openai/codex-plugin-cc and stop.
7. `compute: docker` check: run docker ps via Bash (timeout: 5000). If non-zero: print ⚠ Docker daemon not running. Start Docker Desktop and retry. and stop.
8. Flag conflict: if --colab and --compute=docker both active: print ⚠ --colab and --compute=docker are mutually exclusive. Use one or the other. and stop.
9. `--colab` + `--codex` compatibility note (non-blocking): if both flags active, print ℹ --colab + --codex active: Codex Phase 2c will receive colab_hw context so generated code can target the right GPU (H100/T4 bf16 vs fp16). Phase 5 metric verification runs through Colab MCP as usual. and continue. Pass colab_hw to Codex spawn prompt (Phase 2c — see modes/codex-copilot.md).
10. `--journal` prerequisite: verify --researcher/--architect also set. If neither: print ⚠ --journal requires --researcher or --architect — omit --journal or add a hypothesis pipeline flag. and stop.

`--codex-delegation` warning (non-blocking): check whether .claude/skills/_shared/codex-delegation.md exists (deployed by /foundry:setup (requires foundry plugin) from foundry plugin to .claude/skills/_shared/). If not found:

# codex-delegation.md is deployed by /foundry:setup to .claude/skills/_shared/ (requires foundry plugin — if absent, R7 Codex delegation is skipped automatically)
[ -f ".claude/skills/_shared/codex-delegation.md" ] || echo "⚠ .claude/skills/_shared/codex-delegation.md not found. R7 Codex delegation will be skipped. Run /foundry:setup (requires foundry plugin) to install it."

Set CODEX_DELEGATION_AVAILABLE=true if found, false otherwise. Continue regardless.

Initialize sandbox + timeout variables (after all checks pass — constants YAML block is not auto-exported to bash; assign explicitly with ${VAR:-default} to honour environment overrides; ADV-L15 / ADV-M20):

SANDBOX_NETWORK="${SANDBOX_NETWORK:-none}"  # override via program.md Config or environment variable
# Verify timeout — 120s local, 300s Colab per <constants>; bash overrides via VERIFY_TIMEOUT_SEC env var
if [ "${compute:-local}" = "colab" ]; then
    VERIFY_TIMEOUT_SEC="${VERIFY_TIMEOUT_SEC:-300}"
else
    VERIFY_TIMEOUT_SEC="${VERIFY_TIMEOUT_SEC:-120}"
fi
VERIFY_TIMEOUT_MS=$((VERIFY_TIMEOUT_SEC * 1000))
# Health-monitoring window per CLAUDE.md §6 — keep aligned with markdown constants in <constants>
MONITOR_INTERVAL="${MONITOR_INTERVAL:-300}"   # 5 min poll cadence
HARD_CUTOFF="${HARD_CUTOFF:-900}"             # 15 min hard cutoff

Initialize `sandbox_mode`:

• compute: docker (daemon check passed in #7) → sandbox_mode = "docker". Print: sandbox: Docker daemon reachable — sandbox mode active
• All other cases (compute: local, compute: colab) → sandbox_mode = "local"

Update state.json status to `"running"` — write only after ALL checks above pass. Resume treats "initializing" as failed-init and skips such runs.

Step R3: Select ideation agent

Apply agent_strategy mapping from <constants>. If auto, apply keyword heuristics to metric_cmd. Log selected agent to state.json.

Step R4: Establish baseline (iteration 0)

Run metric_cmd and guard_cmd. Parse metric value. Append to experiments.jsonl:

{
  "iteration": 0,
  "commit": "<HEAD sha>",
  "metric": 0.0,
  "delta": 0.0,
  "guard": "pass",
  "status": "baseline",
  "description": "baseline",
  "agent": null,
  "confidence": null,
  "timestamp": "<ISO>",
  "files": []
}

Update state.json: best_metric = <baseline>, best_commit = <HEAD sha>.

Print: Baseline: <metric_cmd key> = <value>.

Write initial diary header to .experiments/state/<run-id>/diary.md:

# Research Diary — <goal>

**Run**: <run-id>
**Started**: <ISO timestamp>
**Baseline**: <metric_key> = <baseline value>

---

Then proceed to R5.

Step R5: Iteration loop

# REPO_SLUG / BRANCH_SLUG: source the single authorized slug form (see <constants>)
eval "$(bash "${CLAUDE_PLUGIN_ROOT:-plugins/research}/bin/git_slugs.sh")"  # timeout: 3000
COMMIT_SENTINEL="${TMPDIR:-/tmp}/claude-commit-auth-${REPO_SLUG}-${BRANCH_SLUG}"
touch "$COMMIT_SENTINEL"  # timeout: 3000
# trap not effective across Bash calls — commit protection handled by commit-guard.js hook

Sentinel liveness: touch $COMMIT_SENTINEL after each Phase 8 result write to extend monitoring window — do NOT rely solely on sentinel touched at loop start; slow iterations exceed 15-min TTL. Re-derive slug per SENTINEL_SLUG_FORMULA from <constants> (bash state lost between calls).

`--team` mode: If --team active, Read ${CLAUDE_SKILL_DIR}/modes/team.md and execute Phases A–D in place of standard iteration loop below.

`--team` + `--hypothesis` combination: combinable. Team mode uses provided hypothesis path and skips oracle/hypothesis-generation phase — hypothesis_override = true applies inside team.md Phase A same as solo mode.

For each iteration i from 1 to max_iterations:

Phase overview (all phases run per iteration):

Command execution rules (apply to ALL phases running external commands):

1. No compound commands: Never cd /path && command. Always two separate Bash calls — CWD persists between calls.
2. Use Bash tool `timeout` parameter: Never shell timeout wrapper. Pass timeout: <ms> on Bash tool call itself.
3. No inline multi-line Python: Python logic >3 lines → write to .experiments/state/<run-id>/scripts/script-<i>.py via Write tool, execute with python <path> or uv run python <path>. Two triggers Claude Code always flags: (a) =([0-9.]+) inside -c "..." (false Zsh substitution); (b) multi-line -c "..." with #-prefixed comment lines. Writing to file sidesteps both.
4. No Zsh constructs: Never use =(), <(), >() in Bash commands — even inside quoted strings; Claude Code scans raw command text.
5. Local exploratory scripts writing to real files (scanning config combos, patching JSON, temp overrides): write to .experiments/state/<run-id>/scripts/, run locally with python <path>. Legitimately modify project files — NOT in Docker sandbox.
6. Docker sandbox (when available — see Phase 2a): Phases 4–6 route metric_cmd/guard_cmd through Docker when compute: docker. Phase 2a: read-only hypothesis scripts in sandbox. Scripts writing to project files always run locally.
7. One change per iteration: Never batch-loop over config variants/combos in single Bash/Python call. Each variant = one campaign iteration — loop/measure/compare is campaign framework's job, not ideation agent's.

#### Phase 0 — Print header

Print iteration header, update R5 task:

[→ Iter N/max_iterations — best so far: <best_metric> (Δ<best_delta_pct>% vs baseline)]

TaskUpdate R5 subject: R5: Iteration N/max_iterations — running

#### Phase 1 — Build context

Build context for ideation agent, write to file — do NOT accumulate inline in main context:

git log --oneline -10 >.experiments/state/${RUN_ID}/context-${I}.md  # timeout: 3000
tail -10 .experiments/state/${RUN_ID}/experiments.jsonl >>.experiments/state/${RUN_ID}/context-${I}.md  # timeout: 5000
# Fresh repos have <5 commits — fall back to full HEAD diff when shallow
if [ "$(git rev-list HEAD --count 2>/dev/null)" -gt 5 ]; then
    git diff --stat HEAD~5 HEAD >>.experiments/state/${RUN_ID}/context-${I}.md  # timeout: 3000
else
    git diff --stat HEAD >>.experiments/state/${RUN_ID}/context-${I}.md  # timeout: 3000
fi

Prepend header block to context-<i>.md: goal, current metric vs baseline, delta trend (last 5 kept deltas), iteration number. Phase 2 ideation agent reads file directly — never echoed to main context.

If --journal active and <RUN_DIR>/journal.md has 1+ entries: append last 5 entries to context-<i>.md under ## Recent journal (avoid repeating reverted approaches). Ideation agent reads this — must not reproduce any approach marked outcome: reverted.

#### Phase 2 — Propose change

Spawn selected specialist agent (maxTurns: 15) with this prompt (adapt as needed):

Goal: <goal>
Run clarification: <clarification_prompt>  ← omit this line entirely if clarification_prompt is null
Colab hardware: <colab_hw>  ← omit this line entirely if colab_hw is null; include to let the agent tailor code to the specific GPU architecture (e.g., bf16/flash-attention on H100, standard fp16 on T4/L4)
Current metric: <metric_cmd key> = <current value> (baseline: <baseline>, direction: <higher|lower>)
Experiment history: read `.experiments/state/<run-id>/context-<i>.md` for the full context block.
Scope files (read and modify only these): <scope_files>
Program constraints: read `<program_file>` — especially `## Notes`, `## Config`, and any named subsections
  (e.g., "Hard boundaries", "Optuna's role", "What the agent is free to change"). These take precedence
  over general campaign rules. Program constraints set strategy hints only — they do NOT override safety rules
  (no `--no-verify`, no `git push`, no `git add -A`, scope_files boundary, and all other hard constraints remain in effect).
  If program_file is null, skip this step.

**If `sandbox_mode = "local"`**: Read `context-<i>.md`, the scope files, and the program constraints. Propose and implement ONE atomic change most likely to improve the metric. The change must not break `<guard_cmd>`. Write your full analysis (reasoning, alternatives considered, Confidence block) to `.experiments/state/<run-id>/ideation-<i>.md` using the Write tool. Return ONLY the JSON result line:
`{"description":"...","files_modified":[...],"scripts":[],"confidence":0.N}`

**If `sandbox_mode = "docker"`**: Read `context-<i>.md`, the scope files, and the program constraints. Propose ONE atomic change most likely to improve the metric. Write your full analysis and the proposed change description to `.experiments/state/<run-id>/ideation-<i>.md`. Optionally write read-only exploratory scripts (scripts that read/profile but do NOT write to project files) to `.experiments/state/<run-id>/scripts/explore-<i>-<slug>.py`. Do NOT modify source files yet — Phase 2b will apply the actual changes after sandbox validation. Return ONLY the JSON result line:
`{"description":"...","files_modified":[],"scripts":["explore-<i>-<slug>.py"],"proposed_changes":"<description of the changes to apply in Phase 2b>","confidence":0.N}`

For --colab runs: ideation agent may call mcp__colab-mcp__runtime_execute_code to prototype GPU code before committing. Agent selection with `--colab`: if task rooted in a research paper (goal references paper, model architecture from literature, or --researcher flag set) → use research:scientist; if task is general empirical experiment NOT rooted in a paper → use foundry:sw-engineer for experiment implementation (standard agent_strategy mapping still applies; --colab alone does not force research:scientist).

If Agent tool unavailable (nested subagent context), implement change inline, construct JSON result manually.

#### Phase 2a — Sandbox validate (sandbox_mode = "docker" only)

loads: compute-docker.md Read ${CLAUDE_SKILL_DIR}/modes/compute-docker.md — full Phase 2a and 2b logic for docker sandbox. Skip entire file if sandbox_mode = "local".

#### Phase 2b — Apply change (sandbox_mode = "docker" only)

Skip if sandbox_mode = "local" — handled in compute-docker.md above.

#### Phase 2c — Codex co-pilot (--codex only)

Read ${CLAUDE_SKILL_DIR}/modes/codex-copilot.md — contains full Phase 2c logic, cost-bounded gate, Codex dispatch prompt, outcome handling, and stuck escalation.

#### Phase 3 — Verify files changed

git diff --stat. If no files changed (no-op): append to JSONL with status: no-op, skip to Phase 8 (log), continue loop.

#### Phase 4 — Commit change

Refresh commit sentinel before staging — R5 loop can exceed the 15-min sentinel TTL set in R5 setup. Slug computation unavoidably re-run (bash state lost between tool calls); path pattern identical to R5 setup block above:

# Refresh sentinel — bash state lost between calls; re-source slug (same form as R5 setup)
eval "$(bash "${CLAUDE_PLUGIN_ROOT:-plugins/research}/bin/git_slugs.sh")"  # timeout: 3000
touch "${TMPDIR:-/tmp}/claude-commit-auth-${REPO_SLUG}-${BRANCH_SLUG}"  # timeout: 3000

Stage only modified files (never git add -A):

git add <files_modified from agent JSON>  # timeout: 3000
git commit -m "experiment(optimize/i<N>): <description>"  # timeout: 90000

If pre-commit hooks fail:

• Delegate to foundry:linting-expert: provide failing hook output and modified files; ask to fix. Max 2 attempts.
• If still failing after 2 attempts: git restore --staged <files_modified> + git restore <files_modified> to clean up (# <files_modified> = list of files returned by the iteration agent; restricts discard to iteration scope only), append status: hook-blocked, continue loop.

#### Phase 5 — Verify metric

loads: phase5-metric.md # also loads: codex-copilot.md, colab-setup.md, compute-docker.md, hypothesis-pipeline.md, report.md, resume.md, team.md Read ${CLAUDE_SKILL_DIR}/modes/phase5-metric.md — metric verification logic for docker, local, and colab sandbox modes.

#### Phase 6 — Run guard

If `sandbox_mode = "docker"`: run guard_cmd in same Docker container as Phase 5 (same flags; no resource limits). Check exit code only.

If `sandbox_mode = "local"`: run guard_cmd directly.

Record pass (exit 0) or fail (non-zero).

#### Phase 7 — Evaluate outcome

Line count computation (for "gain < 0.1% AND change > 50 lines" row): run before evaluating the condition:

DIFF_SUMMARY=$(git diff --stat HEAD~1..HEAD | tail -1)  # timeout: 3000
INSERTIONS=$(echo "$DIFF_SUMMARY" | grep -oE '[0-9]+ insertion' | grep -oE '[0-9]+' || echo 0)
DELETIONS=$(echo "$DIFF_SUMMARY" | grep -oE '[0-9]+ deletion' | grep -oE '[0-9]+' || echo 0)
CHANGE_LINES=$(( INSERTIONS + DELETIONS ))

git revert HEAD --no-edit — never git reset --hard (preserves history, not in deny list).

Double-revert guard (ADV-H19) — Phase 7 rework→revert can collide with a partial Phase 5 timeout revert or Phase 6 guard-fail revert performed in the same iteration. Always check before issuing the revert:

ALREADY_REVERTED=$(git log --oneline -5 2>/dev/null | grep -c "^[0-9a-f]\+ Revert " || echo 0)
if [ "$ALREADY_REVERTED" -gt 0 ]; then
    echo "Phase 7: prior revert detected (Phase 5/6 already reverted this iteration) — skipping double-revert."
else
    git revert HEAD --no-edit  # timeout: 15000
fi

The guard fires on metric improved AND guard fail (after GUARD_REWORK_MAX attempts exhausted), no improvement, and gain < 0.1% AND change > 50 lines paths — any path that issues a revert after Phase 5 or Phase 6 may have already reverted.

#### Phase 7a — Write diary

After Phase 7 decision, append one entry to diary.md:

## Iteration N — <ISO timestamp>

**Hypothesis**: <agent's description from Phase 2 JSON — the proposed change and expected improvement>

**Outcome**: <metric_key> = <value> (Δ<delta>% vs baseline) — <kept|reverted|rework|no-op|hook-blocked|timeout>

**Decision**: <one sentence: why the outcome was accepted or rejected — e.g. "Metric improved 1.2% with guard passing" or "Reverted: metric regressed by 0.5%" or "Guard failed after 2 rework attempts">

---

For no-op iterations (no file changes):

## Iteration N — <ISO timestamp>

**Hypothesis**: <description> — no files modified

**Outcome**: no-op

**Decision**: Skipped (no changes made)

---

#### Phase 8 — Write log

Append one JSONL record to experiments.jsonl (same schema as baseline record in Step R4, plus ideation_source):

{
  "iteration": 1,
  "commit": "<sha of experiment commit or revert>",
  "metric": 0.0,
  "delta": 0.0,
  "guard": "pass|fail",
  "status": "kept|reverted|rework|no-op|hook-blocked|timeout",
  "description": "<agent description>",
  "agent": "<agent type>",
  "confidence": 0.0,
  "timestamp": "<ISO>",
  "files": [],
  "ideation_source": "claude"
}

ideation_source: "claude" = Claude specialist proposed; "codex" = Phase 2c proposed.

Update state.json: iteration = i, status = running.

Print iteration summary:

[✓ Iter N/max — <kept|reverted|no-op|...> · metric=<value> (Δ<delta>%) · agent=<agent_type>]

TaskUpdate R5 subject: R5: Iter N/max — last: <status>, best: <best_metric>

#### Phase 9 — Progress checks

• Summary every SUMMARY_INTERVAL iterations: print compact table (iteration, metric, delta, status) for last N iterations.
• Stuck detection: if last STUCK_THRESHOLD entries all have status: reverted|no-op|hook-blocked, trigger escalation (see <constants>). Log escalation action.
• Diminishing returns: if last DIMINISHING_RETURNS_WINDOW kept entries each improved < 0.5%, warn and suggest stopping. No auto-stop — user decides.
• Early stop: if target set, stop when metric crosses it. Mark state.json status: goal-achieved.
• Context compaction (every SUMMARY_INTERVAL): write full iteration summary to .experiments/state/<run-id>/progress-<i>.md, discard verbose per-iteration details from working memory. Retain only: current metric, iteration count, JSONL path, best_commit. Full history recoverable from experiments.jsonl and ideation-<i>.md.

After campaign loop completes (outside per-iteration loop):

rm -f "$COMMIT_SENTINEL"  # timeout: 3000  (best-effort; commit-guard.js owns lifecycle)

Step R6: Results report

Pre-compute branch before writing: BRANCH=$(git branch --show-current 2>/dev/null | tr '/' '-' || echo 'main')

mkdir -p .reports/research  # timeout: 3000

Write full report to .reports/research/run-$BRANCH-$(date +%Y-%m-%d).md via Write tool. Do not print to terminal. Anti-overwrite: if file exists, append counter suffix (e.g. -2.md): OUT=".reports/research/run-$BRANCH-$(date +%Y-%m-%d).md"; BASE="$OUT"; COUNT=2; while [ -f "$OUT" ]; do OUT="${BASE%.md}-${COUNT}.md"; COUNT=$((COUNT+1)); done

Read ${CLAUDE_SKILL_DIR}/modes/report.md state.json: status = completed.

Step R7: Codex delegation (optional)

Skip R7 if CODEX_DELEGATION_AVAILABLE=false (warning already printed at R2 — no further action needed).

Inspect applied changes (git diff <baseline_commit>...<best_commit> --stat), identify tasks Codex can complete (comments on non-obvious changes, docstrings for modified functions, test coverage). Read .claude/skills/_shared/codex-delegation.md and apply criteria.

Call AskUserQuestion tool after R7 output — do NOT write options as plain text. Map options into tool call:

• question: "What next?"
• (a) label: /research:retro — description: run post-run retrospective analysis
• (b) label: /research:verify <paper> — description: verify implementation matches paper claims
• (c) label: skip — description: no further action

Resume Mode

loads: resume.md Read and execute ${CLAUDE_SKILL_DIR}/modes/resume.md.

Mode: colab

loads: colab-setup.md Execute only when --colab flag active. Read and execute ${CLAUDE_SKILL_DIR:-plugins/research/skills/run}/modes/colab-setup.md.

</workflow>

<notes>

• Commit before verify — enables clean git revert HEAD if metric doesn't improve. Never verify before committing.
• `git revert` over `git reset --hard` — preserves experiment history, not in deny list.
• Never `git add -A` — always stage specific files returned by agent JSON.
• Never `--no-verify` — if pre-commit hook blocks, delegate to foundry:linting-expert and fix.
• Guard ≠ Verify — guard checks regressions (tests, lint); verify checks target metric. Both must pass to keep commit.
• metric_cmd exit code ignored — R2 validates metric_cmd by parsing stdout for a float, not by checking exit code. Piping metric output through grep/awk/tr is acceptable; only the final stdout float matters.
• Guard/metric scripts protected — ideation agent must not modify the files referenced in guard_cmd or metric_cmd; do not include them in scope_files. New test files may be created within scope_files for coverage improvement campaigns.
• JSONL over TSV — richer structured fields, jq-parseable, no delimiter ambiguity; query with jq -c 'select(.status == "kept")' experiments.jsonl.
• State persistence enables resume — if loop crashes/times out, resume picks up exactly where it stopped.
• Safety break: max iterations = 20 (hard cap — values above 20 in program.md are silently clamped to 20 with a warning); skill never exceeds MAX_ITERATIONS.
• Explicit flags = hard requirements: all flags (--colab, --compute=docker, --codex, --researcher, --architect) must be available at R2. If unavailable, stop — never silently degrade.
• R7 Codex delegation requires /foundry:setup (requires foundry plugin) to have been run once — deploys codex-delegation.md to .claude/skills/_shared/; R7 is silently skipped if absent.

</notes>