The Devin MCP server's backend enum only listed mock/claude/codex/copilot,
excluding the handoff backend that was merged to the engine in #125. This
made the subscription-friendly, no-API-key path unavailable to Devin users
while Claude Code had a dedicated /skillopt-sleep-handoff command for it.
Add "handoff" to the backend enum in _TOOL_SCHEMA so sleep_run accepts
backend: "handoff". The engine already handles the prompt/answer file loop
(exit code 3 + .skillopt-sleep-handoff/); the MCP server needs no special
handling — it passes through the engine output showing pending prompts.
Update the README, rules snippet, and test_backends_in_enum accordingly.
Addresses maintainer review on the OpenAI-compatible endpoints PR:
1. CLI: accept --backend azure_openai in skillopt_sleep/__main__.py (the
documented command was rejected by the argparse choices).
2. Example runner: exit with the child's return code so watchdog/supervisors
see a failed sleep run as a failure.
3. Error state: clear last_call_error when a retry recovers; set an explicit
"empty response on all N attempts" diagnostic when every attempt returns
empty text.
4. Security guard: the managed-identity path now refuses to send an Azure AD
bearer token to any endpoint outside *.openai.azure.com /
*.cognitiveservices.azure.com — a custom endpoint requires explicit
AZURE_OPENAI_AUTH_MODE=openai_compatible + API key.
5. Provider-neutral requests: compat mode sends only the standard contract
(max_tokens, default 8192 via SKILLOPT_SLEEP_COMPAT_MAX_TOKENS);
provider-specific body fields are opt-in via SKILLOPT_SLEEP_CHAT_EXTRA_BODY
(JSON) — the deepseek model-name inference is removed.
6. Docs: removed the unimplemented OPTIMIZER_*/TARGET_* env-var claim; added a
configuration reference matching the implementation exactly.
7. Tests: tests/test_azure_openai_compat.py — 17 deterministic no-network
unittest cases covering CLI acceptance, compat-vs-Azure client selection,
endpoint resolution, the credential guard, request kwargs (opt-in extra
body / token cap), retry-success error clearing, empty-response
diagnostics, and runner exit-code propagation.
Re-verified live against DeepSeek (deepseek-v4-pro, openai_compatible mode)
after the rework: client type OpenAI, completion returned, no error state.
The LinearScheduler.known_decay_sequence test now reflects the new
t = (step-1)/(total_steps-1) contract where step 1 returns max_lr.
Updated docstrings for midpoint and early-step tests to match.
All 43 tests now pass against the updated formula.
The qwen_chat backend (the generic OpenAI-compatible client) hardcoded
max_tokens and always sent temperature, so reasoning models behind
OpenAI-compatible gateways (GPT-5.x, Claude Opus 4.8 via Azure/LiteLLM)
would 400.
- Add opt-in QWEN_CHAT_USE_MAX_COMPLETION_TOKENS (+ role variants) that
swaps the payload key max_tokens -> max_completion_tokens.
- Treat an explicit empty / none / off temperature as "omit" instead of
collapsing to the 0.7 default (via _resolve_temperature).
- Thread both through configure_qwen_chat / _update_config.
- Defaults unchanged; fully backward compatible. Adds 6 tests.
Fixes#127
Co-authored-by: Chirag Singhal <chirag127@users.noreply.github.com>
Adds --backend handoff: the engine runs all deterministic stages and
outsources attempt/judge/reflect to prompt/answer files an interactive
agent session fills between runs (exit 3 = pending batch, re-run to
resume). Deterministic replay + the prompt-hash answer cache make resume
stateless; sentinel detection aborts any call built from unanswered
output so placeholders never reach scores or staging. Session digests
and mined tasks are pinned per night (secret-redacted) so the sessions
answering prompts cannot shift the task set, and LLM mining is routed
through the same handoff files. Ships a /skillopt-sleep-handoff Claude
Code command that answers each prompt in a fresh-context subagent to
protect the held-out gate.
Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
* fix(skillopt-sleep): surface Claude CLI spawn failures instead of silent zero scores
In _call and attempt_with_tools, the bare 'except Exception: return ""'
swallows FileNotFoundError (e.g. bare 'claude' on Windows with npm .cmd
shim) and any other spawn failure, returning an empty string that the
trainer treats as a legitimate model response that scores 0.0 everywhere.
Now the exception is caught explicitly: last_call_error is set, a
warning is logged, and the empty-string return is preserved for
backward compatibility of the control flow. This mirrors the pattern
from #92 (codex backend) which fixed the same class of 'dead CLI
masquerades as nothing to learn' bug.
Issue: #121
* test(skillopt-sleep): add tests verifying Claude CLI spawn failures are surfaced
Add two tests to TestClaudeCliBackendBare:
- test_spawn_failure_sets_last_call_error: _call sets last_call_error
and returns '' when subprocess.run raises FileNotFoundError.
- test_attempt_tools_spawn_failure_sets_last_call_error: same for
attempt_with_tools.
These prove the fix from the parent commit (surface spawn failures
instead of silently scoring 0) and guard against regressions.
Add explicit assertions for held-out scores and gate actions to the verifier discipline test suite to strengthen its guarantees.
- Assert the concrete held-out baseline and candidate scores in test_gate_rejects_reward_hacking_edit.
- Add test_gate_accepts_beneficial_edit using MockBeneficialBackend to provide a paired case where an edit genuinely improves the held-out slice, expecting accepted=True and gate_action='accept_new_best'.
The Claude Code and Codex plugin shells' runner (run-sleep.sh) only
resolved the engine by searching for a skillopt_sleep/ source directory
on disk. After v0.2.0 shipped the engine on PyPI (skillopt-sleep CLI),
the plugins still errored for users who installed via pip/uv without
cloning the repo — even though the error message said "pip install
skillopt" would work.
Add two fallbacks before the error:
1. skillopt-sleep CLI on PATH (covers uv tool install, pipx, pip install)
2. python -m skillopt_sleep when importable (covers pip install into
the active Python)
Fallback 1 is checked before fallback 2 because uv tool install / pipx
isolate the package from the system Python's import path, so the import
check would fail even though the CLI is available.
The existing source-checkout resolution path is unchanged — repo-clone
users keep working as before. The fix lands in both plugins/run-sleep.sh
(shared, used by Codex) and plugins/claude-code/scripts/run-sleep.sh
(bundled copy, byte-identical, used by the Claude Code marketplace
install which fetches only the plugins/claude-code/ subtree).
Runnable SearchQA splits should remain disjoint. A duplicate manifest id across train, val, or test previously collapsed into the wanted-id set and reused the same row in multiple output splits without warning.
Constraint: Preserve manifest order and output schema for valid manifests.
Rejected: Deduplicate automatically | hiding split overlap would make evaluation contamination harder to notice.
Confidence: high
Scope-risk: narrow
Reversibility: clean
Directive: Fail fast on split-manifest integrity problems instead of repairing them silently.
Tested: uv run --with pytest pytest tests/test_materialize_searchqa.py -q
Tested: uv run --with ruff ruff check scripts/materialize_searchqa.py tests/test_materialize_searchqa.py
Not-tested: Loading the live Hugging Face dataset.
Rollout hard scores can be continuous when smoothed rewards are used. Converting the field through int() turned values like 0.75 into 0, which loses signal before scoring and serialization.
Constraint: Keep the existing RolloutResult dictionary shape unchanged.
Rejected: Clamp hard scores to 0 or 1 | contradicts existing continuous-score support in compute_score and sleep replay types.
Confidence: high
Scope-risk: narrow
Reversibility: clean
Directive: Treat hard as numeric reward data, not only a binary label.
Tested: uv run --with pytest pytest tests/test_types.py tests/test_scoring.py -q
Tested: uv run --with ruff ruff check skillopt/types.py tests/test_types.py
Not-tested: Full benchmark rollouts.
LLM responses can include bracketed prose before the actual JSON array. The old greedy regex spanned from the first '[' to the last ']', causing valid single-array answers to be dropped and making multiple arrays indistinguishable.
Constraint: Keep object extraction behavior unchanged.
Rejected: Return first parseable array | silently guesses when a response has multiple valid arrays.
Confidence: high
Scope-risk: narrow
Reversibility: clean
Directive: Keep array extraction conservative when more than one valid top-level array is present.
Tested: uv run --with pytest pytest tests/test_json_utils.py -q
Tested: uv run --with ruff ruff check skillopt/utils/json_utils.py tests/test_json_utils.py
Not-tested: Full benchmark rollouts.
PR #92 added a per-cycle diagnostics.json that surfaces backend stderr,
optimizer replies, and task responses so a 0.0 night is self-diagnosing.
Those free-text fields can carry credentials (e.g. a codex 401 stderr dump
containing an auth token), so persisting them verbatim was a new on-disk
leak surface.
- Add a shared redact_secrets() in staging.py and route diagnostics.json's
call_error / reflect_raw_head / holdout_detail through it before writing.
- Redact the codex and Claude auth-error log lines too (a secondary sink
when a file log handler is attached); last_call_error stays raw in memory
so _AUTH_MARKERS matching is unaffected.
- Centralize _SECRET_PATTERNS in staging.py (harvest_codex now reuses them)
and extend coverage to AWS / GitHub / Slack / Google / JWT token shapes.
- Tests: secret-shape coverage, private-key blocks, recursive/scalar
passthrough, no over-redaction of plain prose, fail-fast auth-error log
redaction, and an end-to-end check that diagnostics.json has no secret.
Observability-only; the gate and learning algorithm are unchanged.
Co-Authored-By: Claude <noreply@anthropic.com>
Splits CodexCliBackend._call into _call_once + a retry wrapper so transient empties/timeouts are retried instead of silently scored 0, and fails fast on fatal auth/model/version errors (401, refresh_token_reused, token_expired, ChatGPT-account-unsupported, newer-Codex-required). On non-zero exit the CLI error text is surfaced via last_call_error instead of being returned as a model response. Adds per-cycle diagnostics.json (observability only; gate and learning algorithm unchanged) so a 0.0 night self-explains.
Add a regression test to ensure the validation gate correctly rejects
reward-hacking skill edits. It has been observed that optimizers
sometimes propose shortcuts that improve train/replay metrics but fail
to improve held-out behavior. This test codifies that the gate blocks
such artifacts.
Add TestVerifierDiscipline to the test_sleep_engine.py suite:
- Create MockRewardHackingBackend that simulates a reward-hacking rule
which passes the train set but degrades the held-out tasks.
- Assert that the proposed edit is rejected by the gate.
Follow-up from a fresh-context review of the prior commit: CodexCliBackend.attempt_with_tools
(the rollout path for tool-requiring tasks) ran codex exec inline, swallowed all exceptions,
and never set last_call_error — so an auth/model/version failure on the tool path still
produced a silent empty->0 with no diagnostic signal, the exact failure class the prior commit
fixed for the _call path. Now it surfaces timeout/exception/non-zero-exit via last_call_error
(response stays empty; never leaks the CLI error text), so a failed tool rollout shows up in
diagnostics.json. Adds a regression test.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
A nightly sleep cycle could run for weeks emitting held-out 0.0 -> 0.0 (gate reject, zero
edits), indistinguishable from "nothing to learn", when the real cause was the codex backend
returning an error (expired auth / model unsupported on the account / outdated CLI) that got
scored as a failed rollout.
backend (CodexCliBackend):
- split _call into _call_once + a retry wrapper: transient empties/timeouts are retried
instead of silently returning "" (mirrors AzureOpenAIBackend's guard);
- on a non-zero exit, surface the reason via last_call_error and return "" rather than
leaking the CLI error text as if it were a model response;
- fail fast (no retries) on fatal auth/model/version errors (401, refresh_token_reused,
token_expired, "not supported when using Codex with a ChatGPT account",
"requires a newer version of Codex").
backend (CliBackend.reflect): retain last_reflect_raw so a no-edits night is diagnosable.
consolidate: ConsolidationResult now carries per-task held-out detail (response, hard/soft,
fail_reason) + reflect_raw + call_error.
cycle: write diagnostics.json per cycle so a 0.0 night self-explains instead of being a black box.
tests: 4 new (retry-not-silent-zero, auth-error-surfaced-not-scored, holdout-detail, reflect-raw).
Also gitignore the .skillopt-sleep/ runtime dir.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Mirror the copilot MCP server: same rich _TOOL_SCHEMA (source, model,
tasks_file, target_skill_path, max_sessions, max_tasks, lookback_hours,
auto_adopt, json, edit_budget, hour, minute) and generic flag forwarding, plus
sleep_schedule / sleep_unschedule. Devin specifics retained: the ATIF-v1.7
harvest step (run before data-reading actions, engine pointed at it via
--claude-home, default --source claude) and post-adopt sync into .devin/skills/.
Tests + README + rules snippet updated for the 7-tool interface.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Review fixes:
- Path bug: SKILLOPT_DEVIN_CLAUDE_HOME (and SKILLOPT_SLEEP_REPO) read from the
env are now wrapped in os.path.expanduser, so the documented "~/..." config
no longer passes a literal ~ to --claude-home (which yielded zero mined
sessions). expanduser on an absolute default is a no-op.
- tests/test_devin_plugin.py: tool-schema completeness, action→subcommand map,
backend enum, the CLAUDE_HOME expansion regression, and an ATIF-v1.7 harvest
shape test against a bundled fixture.
- plugins/devin/fixtures/devin_sample.json: sample ATIF-v1.7 transcript.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Follow-up to the string-aware brace scan: that change only skipped
double-quoted prose, so brace-shaped text in single quotes, backticks, or
bare prose (e.g. `{op: delete}`, '{x: 1}') still reached json_repair and was
fabricated into a bogus dict — strictly worse than None, since extract_json
feeds the optimizer's skill edits.
Add a _looks_json_like() guard before repair: a genuine JSON object's first
non-space char after `{` is `"` (a key) or `}` (empty). Prose pseudo-objects
start with a bare word and are rejected, while legitimate repair targets
(trailing commas, unescaped quotes inside string values) all begin with `"`
and pass — including objects whose string VALUES contain single quotes or
backticks, which must not be rejected.
Found by an independent GPT-5.5 re-review of the merged #79 code. Adds
regression tests for single-quoted / backticked / bare prose (-> None) and
for legitimate objects with quote/backtick string values (still repaired).
Tests: 30 pass (+3 skip) without json_repair, 33 pass with it, both clean
under -W error::RuntimeWarning.
Co-authored-by: Claude <noreply@anthropic.com>
* Robustness for the claude/codex backends on Windows: argv overflow, subprocess encoding, tolerant JSON, test-eval dirs
Fixes surfaced running SkillOpt end-to-end on the bundled `claude` backend
(local Claude CLI) on Windows. None changes the OpenAI/GPT happy path.
1. skillopt/engine/trainer.py — the final test-eval directory
(test_eval_final/) is written to before being created; add
os.makedirs(..., exist_ok=True), matching the two sibling test-eval dirs.
Without it, summary.json raises FileNotFoundError when a rollout yields
zero predictions.
2. skillopt/model/claude_backend.py
a. Pass the prompt via stdin (not argv): on Windows the whole command line
is capped at ~32 KB and a large optimizer prompt (the success-analyst
minibatch carrying several report trajectories) overflows it with
[WinError 206], killing the run after retries.
b. Pass the system prompt via --append-system-prompt-file (a temp file),
not argv. The system prompt here is the skill being optimized, which
SkillOpt grows over training; since the ~32 KB cap applies to the SUM of
all argv, a grown skill would re-hit [WinError 206] even with the prompt
on stdin.
c. Pin the subprocess encoding to utf-8 (errors="replace"). With text=True
and no encoding=, stdin is encoded with the system codepage; on a zh-CN
box (cp936/GBK) a prompt containing an emoji or some Latin-1 characters
raises UnicodeEncodeError before the CLI even starts, failing every retry.
3. skillopt/model/codex_backend.py — the same utf-8 encoding pin on its
subprocess.run(input=...) call (identical unpinned-encoding pattern).
4. skillopt/utils/json_utils.py — extract_json() returned None for valid-
looking JSON that strict json.loads rejects (unescaped ASCII quotes inside
CJK string values, trailing commas), silently dropping the analyst's edits
on non-schema backends (Claude/Qwen): reflect produces N edits, 0 applied.
Add a json_repair fallback, but only on a single unambiguous object — a
balanced-brace extractor plus a refuse-on-multiple-objects guard — so a
chain-of-thought "scratch + final" response can't make repair silently
return the wrong (discarded) object, which would be worse than None (None is
detectable and retryable; a wrong-but-valid edit is applied blind). Declare
json_repair in requirements.txt and the claude/qwen optional extras so the
fallback is actually present (it otherwise no-ops, dropping edits silently).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
(cherry picked from commit dca74a683e)
* fix(json_utils): harden tolerant JSON fallback from PR #77
Follow-up fixes on top of the cherry-picked Windows-robustness change:
1. Make _top_level_brace_objects() fully string-aware in its OUTER scan, not
just inside an object. A '{' inside quoted prose (e.g. '"set it to {x}"')
no longer starts a candidate object, so extract_json() returns None for
prose pseudo-JSON instead of repairing it into a bogus dict — which would
be strictly worse than dropping the edit, since extract_json feeds the
optimizer's skill edits.
2. Pick the repair candidate BEFORE importing json_repair, so the missing-
dependency RuntimeWarning only fires when there is genuinely a single
malformed object that could have been repaired. Ordinary no-JSON / prose
replies (the common case) now return None silently instead of warning on
every call.
3. Resolve dependency-metadata inconsistency: json_repair is optional, so add
it to the `all` extra (it was already in `claude`/`qwen`) and demote it
from a hard requirement to an optional/commented entry in requirements.txt,
matching the project's convention for backend-specific deps.
Adds regression tests for prose-with-braces (-> None), no-warning-on-plain-
text, single-object repair, and multi-object ambiguity. Existing 22 json
tests still pass with and without json_repair installed.
Co-Authored-By: Claude <noreply@anthropic.com>
---------
Co-authored-by: samuelgoofus-boop <260247789+samuelgoofus-boop@users.noreply.github.com>
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
ClaudeCliBackend._call() and attempt_with_tools() hardcoded --bare,
which skips Claude CLI's credential resolution. This broke subscription-
token auth: every model call silently returned "Not logged in" and
scored 0 — the user saw "baseline 0.0 → candidate 0.0, gate reject"
with no indication of an auth failure.
Fix: only pass --bare when ANTHROPIC_API_KEY is set. The remaining
isolation flags (--disable-slash-commands, --disallowedTools,
--exclude-dynamic-system-prompt-sections, clean temp cwd) already
provide the needed isolation without --bare.
Also adds _detect_cli_error() to log a warning when CLI output matches
known auth error patterns, so auth failures surface loudly instead of
deflating every score to 0.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Yif-Yang