Keep completed lanes from ending on mushy stop summaries

The next repo-local sweep target was ROADMAP #69: completed lane
runs could persist vague control text like “commit push everyting,
keep sweeping $ralph”, which made downstream stop summaries
operationally useless. The fix adds a lane-finished quality floor
that preserves strong summaries, rewrites empty/control-only/too-
short-without-context summaries into a contextual fallback, and
records structured metadata explaining when the fallback fired.

Constraint: Keep legitimate concise lane summaries intact while improving only low-signal completions
Rejected: Blanket-rewrite every completed summary into a templated sentence | would erase useful model-authored detail from good lane outputs
Confidence: high
Scope-risk: narrow
Reversibility: clean
Directive: If lane-finished summary heuristics change later, update the structured `qualityFloorApplied/rawSummary/reasons/wordCount` contract and its regression tests together
Tested: cargo fmt --all --check; cargo clippy --workspace --all-targets -- -D warnings; cargo test --workspace; architect review APPROVE
Not-tested: External OMX consumers that may still ignore the new lane.finished data payload
This commit is contained in:
Yeachan-Heo 2026-04-12 03:22:32 +00:00
parent 17e21bc4ad
commit 26b89e583f
2 changed files with 209 additions and 9 deletions

View File

@ -496,7 +496,7 @@ Model name prefix now wins unconditionally over env-var presence. Regression tes
62. **Worker state file surface not implemented****done (verified 2026-04-12):** current `main` already wires `emit_state_file(worker)` into the worker transition path in `rust/crates/runtime/src/worker_boot.rs`, atomically writes `.claw/worker-state.json`, and exposes the documented reader surface through `claw state` / `claw state --output-format json` in `rust/crates/rusty-claude-cli/src/main.rs`. Fresh proof exists in `runtime` regression `emit_state_file_writes_worker_status_on_transition`, the end-to-end `tools` regression `recovery_loop_state_file_reflects_transitions`, and direct CLI parsing coverage for `state` / `state --output-format json`. Source: Jobdori dogfood.
**Scope note (verified 2026-04-12):** ROADMAP #31, #43, and #63-#68 currently appear to describe acpx/droid or upstream OMX/server orchestration behavior, not claw-code source already present in this repository. Repo-local searches for `acpx`, `use-droid`, `run-acpx`, `commit-wrapper`, `ultraclaw`, `roadmap-nudge-10min`, `OMX_TMUX_INJECT`, `/hooks/health`, and `/hooks/status` found no implementation hits outside `ROADMAP.md`, and the earlier state-surface note already records that the HTTP server is not owned by claw-code. With #45 now fixed, the remaining unresolved items in this section look like external tracking notes rather than confirmed repo-local backlog; re-check if new repo-local evidence appears.
**Scope note (verified 2026-04-12):** ROADMAP #31, #43, and #63-#68 currently appear to describe acpx/droid or upstream OMX/server orchestration behavior, not claw-code source already present in this repository. Repo-local searches for `acpx`, `use-droid`, `run-acpx`, `commit-wrapper`, `ultraclaw`, `roadmap-nudge-10min`, `OMX_TMUX_INJECT`, `/hooks/health`, and `/hooks/status` found no implementation hits outside `ROADMAP.md`, and the earlier state-surface note already records that the HTTP server is not owned by claw-code. With #45 and #69 now fixed, the remaining unresolved items in this section look like external tracking notes rather than confirmed repo-local backlog; re-check if new repo-local evidence appears.
63. **Droid session completion semantics broken: code arrives after "status: completed"** — dogfooded 2026-04-12. Ultraclaw droid sessions (use-droid via acpx) report `session.status: completed` before file writes are fully flushed/synced to the working tree. Discovered +410 lines of "late-arriving" droid output that appeared after I had already assessed 8 sessions as "no code produced." This creates false-negative assessments and duplicate work. **Fix shape:** (a) droid agent should only report completion after explicit file-write confirmation (fsync or existence check); (b) or, claw-code should expose a `pending_writes` status that indicates "agent responded, disk flush pending"; (c) lane orchestrators should poll for file changes for N seconds after completion before final assessment. **Blocker:** none. Source: Jobdori ultraclaw dogfood 2026-04-12.
@ -510,6 +510,6 @@ Model name prefix now wins unconditionally over env-var presence. Regression tes
68. **Internal reinjection/resume paths leak opaque control prose** — dogfooded 2026-04-12. OMX lanes stopping with `Continue from current mode state. [OMX_TMUX_INJECT]` expose internal implementation details instead of operator-meaningful state. The event tells us *that* tmux reinjection happened, but not *why* (retry after failure? resume after idle? manual recovery?), *what state was preserved*, or *what the lane was trying to do*. **Fix shape:** recovery/reinject events should emit structured cause like: `resume_after_stop`, `retry_after_tool_failure`, `tmux_reinject_after_idle`, `manual_recovery` plus preserved state / target lane info. Never leak bare internal markers like `[OMX_TMUX_INJECT]` as the primary summary. Blocker: none. Source: gaebal-gajae dogfood analysis 2026-04-12.
69. **Lane stop summaries have no minimum quality floor**dogfooded 2026-04-12. `clawcode-human` session stopped with summary `commit push everyting, keep sweeping $ralph` — vague, typo-ridden, operationally useless. Unlike well-scoped review lanes, this summary regressed to mushy command prose with no outcome clarity. **Fix shape:** (a) enforce minimum stop/result summary standards: what was done (outcome), what was scoped (target), what's next (state); (b) typo/grammar validation; (c) reject summaries that are shorter than N words or contain only control verbs without context. Blocker: none. Source: gaebal-gajae dogfood analysis 2026-04-12.
69. **Lane stop summaries have no minimum quality floor****done (verified 2026-04-12):** completed lane persistence in `rust/crates/tools/src/lib.rs` now normalizes vague/control-only stop summaries into a contextual fallback that includes the lane target and status, while preserving structured metadata about whether the quality floor fired (`qualityFloorApplied`, `rawSummary`, `reasons`, `wordCount`). Regression coverage locks both the pass-through path for good summaries and the fallback path for mushy summaries like `commit push everyting, keep sweeping $ralph`. **Original filing below.**
70. **Install-source ambiguity misleads real users** — community observation 2026-04-12. User treated `claw-code.io` as official, then hit `clawcode` vs deprecated `claw-code` naming collision and concluded install story was inconsistent. Source-of-truth is not obvious when website/repo/crates naming diverges. **Fix shape:** canonical repo docs should explicitly state which site is official; installation guidance should visibly warn against deprecated `claw-code` crate and ambiguous third-party pages. Blocker: none. Source: gaebal-gajae community watch 2026-04-12.
70. **Install-source ambiguity misleads real users** — community observation 2026-04-12. User treated `claw-code.io` as official, then hit `clawcode` vs deprecated `claw-code` naming collision and concluded install story was inconsistent. Source-of-truth is not obvious when website/repo/crates naming diverges. **Fix shape:** canonical repo docs should explicitly state which site is official; installation guidance should visibly warn against deprecated `claw-code` crate and ambiguous third-party pages. Blocker: none. Source: gaebal-gajae community watch 2026-04-12.

View File

@ -3743,12 +3743,13 @@ fn persist_agent_terminal_state(
.push(LaneEvent::failed(iso8601_now(), &blocker));
} else {
next_manifest.current_blocker = None;
let compressed_detail = result
.filter(|value| !value.trim().is_empty())
.map(|value| compress_summary_text(value.trim()));
next_manifest
.lane_events
.push(LaneEvent::finished(iso8601_now(), compressed_detail));
let finished_summary = build_lane_finished_summary(&next_manifest, result);
next_manifest.lane_events.push(
LaneEvent::finished(iso8601_now(), finished_summary.detail).with_data(
serde_json::to_value(&finished_summary.data)
.expect("lane summary metadata should serialize"),
),
);
if let Some(provenance) = maybe_commit_provenance(result) {
next_manifest.lane_events.push(LaneEvent::commit_created(
iso8601_now(),
@ -3760,6 +3761,152 @@ fn persist_agent_terminal_state(
write_agent_manifest(&next_manifest)
}
const MIN_LANE_SUMMARY_WORDS: usize = 7;
const CONTROL_ONLY_SUMMARY_WORDS: &[&str] = &[
"ack",
"commit",
"continue",
"everyting",
"everything",
"keep",
"next",
"push",
"ralph",
"resume",
"retry",
"run",
"stop",
"sweep",
"sweeping",
"team",
];
const CONTEXTUAL_SUMMARY_WORDS: &[&str] = &[
"added",
"audited",
"blocked",
"completed",
"documented",
"failed",
"finished",
"fixed",
"implemented",
"investigated",
"merged",
"pushed",
"refactored",
"removed",
"reviewed",
"tested",
"updated",
"verified",
];
#[derive(Debug, Clone, Serialize)]
struct LaneFinishedSummaryData {
#[serde(rename = "qualityFloorApplied")]
quality_floor_applied: bool,
reasons: Vec<String>,
#[serde(rename = "rawSummary", skip_serializing_if = "Option::is_none")]
raw_summary: Option<String>,
#[serde(rename = "wordCount")]
word_count: usize,
}
#[derive(Debug, Clone)]
struct LaneFinishedSummary {
detail: Option<String>,
data: LaneFinishedSummaryData,
}
#[derive(Debug)]
struct LaneSummaryAssessment {
apply_quality_floor: bool,
reasons: Vec<String>,
word_count: usize,
}
fn build_lane_finished_summary(
manifest: &AgentOutput,
result: Option<&str>,
) -> LaneFinishedSummary {
let raw_summary = result.map(str::trim).filter(|value| !value.is_empty());
let assessment = assess_lane_summary_quality(raw_summary.unwrap_or_default());
let detail = match raw_summary {
Some(summary) if !assessment.apply_quality_floor => Some(compress_summary_text(summary)),
Some(summary) => Some(compose_lane_summary_fallback(manifest, Some(summary))),
None => Some(compose_lane_summary_fallback(manifest, None)),
};
LaneFinishedSummary {
detail,
data: LaneFinishedSummaryData {
quality_floor_applied: raw_summary.is_none() || assessment.apply_quality_floor,
reasons: assessment.reasons,
raw_summary: raw_summary.map(str::to_string),
word_count: assessment.word_count,
},
}
}
fn assess_lane_summary_quality(summary: &str) -> LaneSummaryAssessment {
let words = summary
.split(|ch: char| !(ch.is_ascii_alphanumeric() || ch == '-' || ch == '#'))
.filter(|token| !token.is_empty())
.map(str::to_ascii_lowercase)
.collect::<Vec<_>>();
let word_count = words.len();
let mut reasons = Vec::new();
if summary.trim().is_empty() {
reasons.push(String::from("empty"));
}
let control_only = !words.is_empty()
&& words
.iter()
.all(|word| CONTROL_ONLY_SUMMARY_WORDS.contains(&word.as_str()));
if control_only {
reasons.push(String::from("control_only"));
}
let has_context_signal = summary.contains('`')
|| summary.contains('/')
|| summary.contains(':')
|| summary.contains('#')
|| words
.iter()
.any(|word| CONTEXTUAL_SUMMARY_WORDS.contains(&word.as_str()));
if word_count < MIN_LANE_SUMMARY_WORDS && !has_context_signal {
reasons.push(String::from("too_short_without_context"));
}
LaneSummaryAssessment {
apply_quality_floor: !reasons.is_empty(),
reasons,
word_count,
}
}
fn compose_lane_summary_fallback(manifest: &AgentOutput, raw_summary: Option<&str>) -> String {
let target = manifest.description.trim();
let base = format!(
"Completed lane `{}` for target: {}. Status: completed.",
manifest.name,
if target.is_empty() {
"unspecified task"
} else {
target
}
);
match raw_summary {
Some(summary) => format!(
"{base} Original stop summary was too vague to keep as the lane result: \"{}\".",
summary.trim()
),
None => format!("{base} No usable stop summary was produced by the lane."),
}
}
fn derive_agent_state(
status: &str,
result: Option<&str>,
@ -7240,6 +7387,14 @@ mod tests {
completed_manifest_json["laneEvents"][1]["event"],
"lane.finished"
);
assert_eq!(
completed_manifest_json["laneEvents"][1]["data"]["qualityFloorApplied"],
false
);
assert_eq!(
completed_manifest_json["laneEvents"][1]["detail"],
"Finished successfully in commit abc1234"
);
assert_eq!(
completed_manifest_json["laneEvents"][2]["event"],
"lane.commit.created"
@ -7301,6 +7456,51 @@ mod tests {
);
assert_eq!(failed_manifest_json["derivedState"], "truly_idle");
let normalized = execute_agent_with_spawn(
AgentInput {
description: "Sweep the next backlog item".to_string(),
prompt: "Produce a low-signal stop summary".to_string(),
subagent_type: Some("Explore".to_string()),
name: Some("summary-floor".to_string()),
model: None,
},
|job| {
persist_agent_terminal_state(
&job.manifest,
"completed",
Some("commit push everyting, keep sweeping $ralph"),
None,
)
},
)
.expect("normalized agent should succeed");
let normalized_manifest = std::fs::read_to_string(&normalized.manifest_file)
.expect("normalized manifest should exist");
let normalized_manifest_json: serde_json::Value =
serde_json::from_str(&normalized_manifest).expect("normalized manifest json");
assert_eq!(
normalized_manifest_json["laneEvents"][1]["event"],
"lane.finished"
);
let normalized_detail = normalized_manifest_json["laneEvents"][1]["detail"]
.as_str()
.expect("normalized detail");
assert!(normalized_detail.contains("Completed lane `summary-floor`"));
assert!(normalized_detail.contains("Sweep the next backlog item"));
assert_eq!(
normalized_manifest_json["laneEvents"][1]["data"]["qualityFloorApplied"],
true
);
assert_eq!(
normalized_manifest_json["laneEvents"][1]["data"]["rawSummary"],
"commit push everyting, keep sweeping $ralph"
);
assert_eq!(
normalized_manifest_json["laneEvents"][1]["data"]["reasons"][0],
"control_only"
);
let spawn_error = execute_agent_with_spawn(
AgentInput {
description: "Spawn error task".to_string(),