mirror of
https://github.com/boshu2/agentops.git
synced 2026-09-14 15:08:13 +08:00
Claude
f02f0856de
The Dream curator emits Source: "dream-degraded" for self-reported overnight degradations (cli/cmd/ao/overnight_packets.go:719), but the next-work v1.4 contract enum did not list it. The pre-push contract parity gate then blocked any push that occurred after a dream pass had populated the live queue — visible as `next-work.jsonl has active item enum drift: source=dream-degraded`. Add `dream-degraded` to: - docs/contracts/next-work.schema.md §Source - scripts/validate-next-work-contract-parity.sh (item_sources + jq valid_source) - scripts/check-next-work-schema-rows.sh (VALID_SOURCES) - skills/post-mortem/references/harvest-next-work.md (case statement) - skills-codex/post-mortem/references/harvest-next-work.md (case statement) Add a parity test fixture that asserts an active dream-degraded item passes the gate, locking producer/consumer reconciliation.
443 lines
16 KiB
Bash
Executable File
443 lines
16 KiB
Bash
Executable File
#!/usr/bin/env bash
|
|
set -euo pipefail
|
|
|
|
SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
|
|
ROOT="$(cd "$SCRIPT_DIR/.." && pwd)"
|
|
|
|
usage() {
|
|
cat <<'EOF'
|
|
Usage: bash scripts/validate-next-work-contract-parity.sh [repo-root]
|
|
EOF
|
|
}
|
|
|
|
if [[ "${1:-}" == "-h" || "${1:-}" == "--help" ]]; then
|
|
usage
|
|
exit 0
|
|
fi
|
|
|
|
if [[ $# -gt 1 ]]; then
|
|
usage >&2
|
|
exit 2
|
|
fi
|
|
|
|
if [[ $# -eq 1 ]]; then
|
|
ROOT="$1"
|
|
fi
|
|
|
|
if [[ "$ROOT" != /* ]]; then
|
|
ROOT="$(cd "$ROOT" && pwd)"
|
|
fi
|
|
|
|
SCHEMA="$ROOT/docs/contracts/next-work.schema.md"
|
|
HARVEST_REF="$ROOT/skills/post-mortem/references/harvest-next-work.md"
|
|
POST_MORTEM_SKILL="$ROOT/skills/post-mortem/SKILL.md"
|
|
POST_MORTEM_CODEX_SKILL="$ROOT/skills-codex/post-mortem/SKILL.md"
|
|
PHASE_CONTRACT="$ROOT/skills/rpi/references/phase-data-contracts.md"
|
|
GATE4="$ROOT/skills/rpi/references/gate4-loop-and-spawn.md"
|
|
RUNTIME="$(mktemp -t nextwork-runtime.XXXXXX)"
|
|
trap 'rm -f "$RUNTIME"' EXIT
|
|
cat \
|
|
"$ROOT/cli/internal/rpi/types.go" \
|
|
"$ROOT/cli/internal/rpi/helpers.go" \
|
|
"$ROOT/cli/cmd/ao/rpi_loop.go" \
|
|
> "$RUNTIME" 2>/dev/null || true
|
|
SMOKE="$ROOT/tests/smoke-test.sh"
|
|
|
|
failures=0
|
|
|
|
fail() {
|
|
echo "FAIL: $1" >&2
|
|
failures=$((failures + 1))
|
|
}
|
|
|
|
have_rg() {
|
|
command -v rg >/dev/null 2>&1
|
|
}
|
|
|
|
require_file() {
|
|
local path="$1"
|
|
[[ -f "$path" ]] || fail "missing file: ${path#$ROOT/}"
|
|
}
|
|
|
|
contains_fixed_file() {
|
|
local needle="$1"
|
|
local path="$2"
|
|
|
|
if have_rg; then
|
|
rg -Fq "$needle" "$path"
|
|
return
|
|
fi
|
|
|
|
grep -Fq -- "$needle" "$path"
|
|
}
|
|
|
|
contains_fixed_stdin() {
|
|
local needle="$1"
|
|
|
|
if have_rg; then
|
|
rg -Fq "$needle"
|
|
return
|
|
fi
|
|
|
|
grep -Fq -- "$needle"
|
|
}
|
|
|
|
require_contains() {
|
|
local path="$1"
|
|
local needle="$2"
|
|
local label="$3"
|
|
if ! contains_fixed_file "$needle" "$path"; then
|
|
fail "$label"
|
|
fi
|
|
}
|
|
|
|
require_section_contains() {
|
|
local path="$1"
|
|
local start="$2"
|
|
local end="$3"
|
|
local needle="$4"
|
|
local label="$5"
|
|
local section
|
|
|
|
section="$(awk -v start="$start" -v end="$end" '
|
|
index($0, start) {in_section=1}
|
|
in_section {print}
|
|
in_section && index($0, end) {exit}
|
|
' "$path")"
|
|
|
|
if [[ -z "$section" ]]; then
|
|
fail "missing section $start in ${path#$ROOT/}"
|
|
return
|
|
fi
|
|
|
|
if ! printf '%s\n' "$section" | contains_fixed_stdin "$needle"; then
|
|
fail "$label"
|
|
fi
|
|
}
|
|
|
|
require_section_not_contains() {
|
|
local path="$1"
|
|
local start="$2"
|
|
local end="$3"
|
|
local needle="$4"
|
|
local label="$5"
|
|
local section
|
|
|
|
section="$(awk -v start="$start" -v end="$end" '
|
|
index($0, start) {in_section=1}
|
|
in_section {print}
|
|
in_section && index($0, end) {exit}
|
|
' "$path")"
|
|
|
|
if [[ -z "$section" ]]; then
|
|
fail "missing section $start in ${path#$ROOT/}"
|
|
return
|
|
fi
|
|
|
|
if printf '%s\n' "$section" | contains_fixed_stdin "$needle"; then
|
|
fail "$label"
|
|
fi
|
|
}
|
|
|
|
for path in \
|
|
"$SCHEMA" \
|
|
"$HARVEST_REF" \
|
|
"$POST_MORTEM_SKILL" \
|
|
"$POST_MORTEM_CODEX_SKILL" \
|
|
"$PHASE_CONTRACT" \
|
|
"$GATE4" \
|
|
"$RUNTIME" \
|
|
"$SMOKE"; do
|
|
require_file "$path"
|
|
done
|
|
|
|
require_contains "$SCHEMA" "schema_version: 1.4" \
|
|
"next-work schema is not at v1.4"
|
|
require_contains "$SCHEMA" 'Item lifecycle inside `items[]` is authoritative.' \
|
|
"next-work schema must declare item lifecycle authority"
|
|
require_contains "$SCHEMA" "may be empty when a post-mortem finds nothing actionable" \
|
|
"next-work schema must permit empty items arrays"
|
|
require_contains "$SCHEMA" "consumers may rewrite existing lines to claim, release, fail, or consume individual items" \
|
|
"next-work schema must describe rewrite semantics"
|
|
require_contains "$SCHEMA" "Legacy Compatibility" \
|
|
"next-work schema must document legacy flat rows"
|
|
|
|
entry_fields=(
|
|
source_epic timestamp items consumed claim_status claimed_by claimed_at
|
|
consumed_by consumed_at failed_at
|
|
)
|
|
item_fields=(
|
|
title type severity source description evidence target_repo proof_ref consumed
|
|
claim_status claimed_by claimed_at consumed_by consumed_at failed_at
|
|
)
|
|
proof_ref_fields=(
|
|
kind target_id run_id path
|
|
)
|
|
proof_ref_kinds=(
|
|
completed_run evidence_only_closure execution_packet
|
|
)
|
|
item_types=(
|
|
tech-debt improvement pattern-fix process-improvement feature bug task docs chore
|
|
)
|
|
item_sources=(
|
|
council-finding retro-learning retro-pattern evolve-generator
|
|
feature-suggestion backlog-processing post-mortem-finding manifest-classification
|
|
dream-degraded
|
|
)
|
|
|
|
for field in "${entry_fields[@]}"; do
|
|
require_contains "$SCHEMA" "\`$field\`" \
|
|
"next-work schema missing entry field \`$field\`"
|
|
require_contains "$RUNTIME" "json:\"$field" \
|
|
"runtime next-work structs missing json field $field"
|
|
done
|
|
|
|
for field in "${item_fields[@]}"; do
|
|
require_contains "$SCHEMA" "\`$field\`" \
|
|
"next-work schema missing item field \`$field\`"
|
|
done
|
|
|
|
for field in "${proof_ref_fields[@]}"; do
|
|
require_contains "$SCHEMA" "\`$field\`" \
|
|
"next-work schema missing proof_ref field \`$field\`"
|
|
done
|
|
|
|
for value in "${proof_ref_kinds[@]}"; do
|
|
require_contains "$SCHEMA" "\`$value\`" \
|
|
"next-work schema missing proof_ref kind $value"
|
|
require_contains "$HARVEST_REF" "$value" \
|
|
"harvest-next-work reference missing proof_ref kind $value"
|
|
done
|
|
|
|
require_contains "$RUNTIME" 'json:"proof_ref,omitempty"' \
|
|
"runtime next-work structs missing json field proof_ref"
|
|
for field in kind target_id run_id path; do
|
|
require_contains "$RUNTIME" "json:\"$field" \
|
|
"runtime proof_ref struct missing json field $field"
|
|
done
|
|
|
|
for value in "${item_types[@]}"; do
|
|
require_contains "$SCHEMA" "\`$value\`" \
|
|
"next-work schema missing type enum value $value"
|
|
require_contains "$HARVEST_REF" "$value" \
|
|
"harvest-next-work reference missing type enum value $value"
|
|
done
|
|
|
|
for value in high medium low; do
|
|
require_contains "$SCHEMA" "\`$value\`" \
|
|
"next-work schema missing severity enum value $value"
|
|
done
|
|
|
|
for value in "${item_sources[@]}"; do
|
|
require_contains "$SCHEMA" "\`$value\`" \
|
|
"next-work schema missing source enum value $value"
|
|
require_contains "$HARVEST_REF" "$value" \
|
|
"harvest-next-work reference missing source enum value $value"
|
|
done
|
|
|
|
for value in available in_progress consumed; do
|
|
require_contains "$SCHEMA" "\`$value\`" \
|
|
"next-work schema missing claim_status value $value"
|
|
require_contains "$RUNTIME" "\"$value\"" \
|
|
"runtime next-work logic missing claim_status value $value"
|
|
done
|
|
|
|
require_contains "$HARVEST_REF" "docs/contracts/next-work.schema.md" \
|
|
"harvest-next-work must reference the tracked next-work schema"
|
|
require_contains "$HARVEST_REF" "proof_ref" \
|
|
"harvest-next-work must document proof_ref emission"
|
|
require_contains "$POST_MORTEM_SKILL" "docs/contracts/next-work.schema.md" \
|
|
"post-mortem skill must reference the tracked next-work schema"
|
|
require_contains "$POST_MORTEM_CODEX_SKILL" "docs/contracts/next-work.schema.md" \
|
|
"generated Codex post-mortem skill must reference the tracked next-work schema"
|
|
require_contains "$GATE4" "docs/contracts/next-work.schema.md" \
|
|
"rpi gate4 reference must point at the tracked next-work schema"
|
|
require_contains "$PHASE_CONTRACT" "item lifecycle as authoritative" \
|
|
"phase-data-contracts must describe item lifecycle authority"
|
|
require_contains "$PHASE_CONTRACT" 'entry aggregate flips to `consumed=true` only after every child item is consumed' \
|
|
"phase-data-contracts must describe aggregate consumption rule"
|
|
require_contains "$GATE4" "Never mark an item consumed at pick-time" \
|
|
"rpi gate4 must retain claim-before-consume rule"
|
|
require_contains "$SMOKE" "validate-next-work-contract-parity.sh" \
|
|
"smoke-test must execute the next-work contract parity validator"
|
|
|
|
require_contains "$RUNTIME" "case \"feature\", \"improvement\", \"tech-debt\", \"pattern-fix\", \"bug\", \"task\":" \
|
|
"RPI runtime is missing workTypeRank coverage for pattern-fix"
|
|
require_contains "$RUNTIME" "case \"process-improvement\", \"docs\", \"chore\":" \
|
|
"RPI runtime is missing workTypeRank coverage for docs/chore"
|
|
require_contains "$RUNTIME" 'omitted item `claim_status` semantically' \
|
|
"runtime comments or docs should preserve omitted claim_status semantics"
|
|
|
|
# Proof-backed completion evidence — runtime must use proof, not heuristic suppression.
|
|
require_contains "$RUNTIME" 'CompletionEvidence' \
|
|
"runtime must reference CompletionEvidence for proof-backed skip logic"
|
|
require_contains "$RUNTIME" 'completion_evidence' \
|
|
"runtime next-work struct must have completion_evidence json tag"
|
|
|
|
# Harvest reference must show the batched v1.4 write shape.
|
|
require_contains "$HARVEST_REF" "source_epic" \
|
|
"harvest-next-work reference must show batched source_epic field"
|
|
require_contains "$HARVEST_REF" "items" \
|
|
"harvest-next-work reference must show batched items array"
|
|
require_contains "$HARVEST_REF" 'claim_status' \
|
|
"harvest-next-work reference must show claim_status lifecycle"
|
|
require_contains "$HARVEST_REF" "Schema Validation" \
|
|
"harvest-next-work reference must include schema validation section"
|
|
|
|
# Crank skill must not contain flat next-work append examples.
|
|
CRANK_SKILL="$ROOT/skills/crank/SKILL.md"
|
|
if [[ -f "$CRANK_SKILL" ]]; then
|
|
if contains_fixed_file 'echo "{\"title\":' "$CRANK_SKILL" 2>/dev/null; then
|
|
fail "crank skill contains legacy flat-row next-work append example"
|
|
fi
|
|
fi
|
|
|
|
for skill in "$POST_MORTEM_SKILL" "$POST_MORTEM_CODEX_SKILL"; do
|
|
require_section_contains "$skill" '#### Step ACT.3: Feed Next-Work' '#### Step ACT.4: Update Marker' \
|
|
"docs/contracts/next-work.schema.md" \
|
|
"${skill#$ROOT/} ACT.3 must reference the tracked next-work schema"
|
|
require_section_contains "$skill" '#### Step ACT.3: Feed Next-Work' '#### Step ACT.4: Update Marker' \
|
|
"source_epic:" \
|
|
"${skill#$ROOT/} ACT.3 must show batched next-work entries"
|
|
require_section_contains "$skill" '#### Step ACT.3: Feed Next-Work' '#### Step ACT.4: Update Marker' \
|
|
"items:" \
|
|
"${skill#$ROOT/} ACT.3 must show batched item arrays"
|
|
require_section_contains "$skill" '#### Step ACT.3: Feed Next-Work' '#### Step ACT.4: Update Marker' \
|
|
"\"proof_ref\"" \
|
|
"${skill#$ROOT/} ACT.3 must show proof_ref emission guidance"
|
|
require_section_contains "$skill" '#### Step ACT.3: Feed Next-Work' '#### Step ACT.4: Update Marker' \
|
|
'claim_status: "available"' \
|
|
"${skill#$ROOT/} ACT.3 must initialize entry claim_status"
|
|
require_section_not_contains "$skill" '#### Step ACT.3: Feed Next-Work' '#### Step ACT.4: Update Marker' \
|
|
'echo "{\"title\":' \
|
|
"${skill#$ROOT/} ACT.3 still contains the legacy flat-row append example"
|
|
done
|
|
|
|
# Drift validation: if a live next-work.jsonl exists, verify all entries conform to v1.4.
|
|
LIVE_QUEUE="$ROOT/.agents/rpi/next-work.jsonl"
|
|
if [[ -f "$LIVE_QUEUE" ]] && command -v jq >/dev/null 2>&1; then
|
|
drift_count=0
|
|
while IFS= read -r line; do
|
|
[[ -z "$line" ]] && continue
|
|
if ! echo "$line" | jq -e '.source_epic and .items and (.claim_status != null)' >/dev/null 2>&1; then
|
|
drift_count=$((drift_count + 1))
|
|
fi
|
|
done < "$LIVE_QUEUE"
|
|
if [[ "$drift_count" -gt 0 ]]; then
|
|
fail "next-work.jsonl has $drift_count entries not conforming to v1.4 schema (missing source_epic, items, or claim_status)"
|
|
fi
|
|
|
|
aggregate_self_drift="$(
|
|
jq -s -r '
|
|
to_entries[] as $line |
|
|
select(($line.value.items? | type) == "array") |
|
|
select(
|
|
(((($line.value.consumed // false) == true) and ($line.value.claim_status != "consumed")) or
|
|
((($line.value.consumed // false) != true) and ($line.value.claim_status == "consumed")))
|
|
) |
|
|
"line \($line.key + 1) source_epic=\($line.value.source_epic // "<missing>") consumed=\($line.value.consumed // false) claim_status=\($line.value.claim_status // "<missing>")"
|
|
' "$LIVE_QUEUE" 2>/dev/null || true
|
|
)"
|
|
if [[ -n "$aggregate_self_drift" ]]; then
|
|
fail "next-work.jsonl has aggregate lifecycle self drift: $(printf '%s\n' "$aggregate_self_drift" | head -1)"
|
|
fi
|
|
|
|
enum_drift="$(
|
|
jq -s -r '
|
|
def item_status:
|
|
if .claim_status == "in_progress" then "in_progress"
|
|
elif ((.consumed // false) == true) or (.claim_status == "consumed") then "consumed"
|
|
else "available"
|
|
end;
|
|
def explicit_lifecycle:
|
|
has("claim_status") or
|
|
((.consumed // false) == true) or
|
|
has("claimed_by") or
|
|
has("claimed_at") or
|
|
has("consumed_by") or
|
|
has("consumed_at") or
|
|
has("failed_at");
|
|
def aggregate_status:
|
|
if ((.consumed // false) == true) or (.claim_status == "consumed") then "consumed"
|
|
elif .claim_status == "in_progress" then "in_progress"
|
|
else "available"
|
|
end;
|
|
def active_item($entry):
|
|
if explicit_lifecycle then item_status != "consumed"
|
|
else ($entry | aggregate_status) != "consumed"
|
|
end;
|
|
def valid_type:
|
|
. == "tech-debt" or . == "improvement" or . == "pattern-fix" or
|
|
. == "process-improvement" or . == "feature" or . == "bug" or . == "task" or
|
|
. == "docs" or . == "chore";
|
|
def valid_severity:
|
|
. == "high" or . == "medium" or . == "low";
|
|
def valid_source:
|
|
. == "council-finding" or . == "retro-learning" or . == "retro-pattern" or
|
|
. == "evolve-generator" or . == "feature-suggestion" or . == "backlog-processing" or
|
|
. == "post-mortem-finding" or . == "manifest-classification" or
|
|
. == "dream-degraded";
|
|
to_entries[] as $line |
|
|
select(($line.value.items? | type) == "array") |
|
|
($line.value.items | to_entries[]) as $item |
|
|
select($item.value | active_item($line.value)) |
|
|
select(
|
|
(($item.value.type | valid_type | not) or
|
|
($item.value.severity | valid_severity | not) or
|
|
($item.value.source | valid_source | not))
|
|
) |
|
|
"line \($line.key + 1) source_epic=\($line.value.source_epic // "<missing>") item=\($item.key + 1) type=\($item.value.type // "<missing>") severity=\($item.value.severity // "<missing>") source=\($item.value.source // "<missing>")"
|
|
' "$LIVE_QUEUE" 2>/dev/null || true
|
|
)"
|
|
if [[ -n "$enum_drift" ]]; then
|
|
fail "next-work.jsonl has active item enum drift: $(printf '%s\n' "$enum_drift" | head -1)"
|
|
fi
|
|
|
|
lifecycle_drift="$(
|
|
jq -s -r '
|
|
def item_status:
|
|
if .claim_status == "in_progress" then "in_progress"
|
|
elif ((.consumed // false) == true) or (.claim_status == "consumed") then "consumed"
|
|
else "available"
|
|
end;
|
|
def explicit_lifecycle:
|
|
has("claim_status") or
|
|
((.consumed // false) == true) or
|
|
has("claimed_by") or
|
|
has("claimed_at") or
|
|
has("consumed_by") or
|
|
has("consumed_at") or
|
|
has("failed_at");
|
|
def aggregate_status:
|
|
if ((.consumed // false) == true) or (.claim_status == "consumed") then "consumed"
|
|
elif .claim_status == "in_progress" then "in_progress"
|
|
else "available"
|
|
end;
|
|
def expected_status($statuses):
|
|
if all($statuses[]; . == "consumed") then "consumed"
|
|
elif any($statuses[]; . == "in_progress") then "in_progress"
|
|
else "available"
|
|
end;
|
|
to_entries[] as $line |
|
|
select(($line.value.items? | type) == "array") |
|
|
($line.value.items | map(select(explicit_lifecycle))) as $explicit |
|
|
select(($explicit | length) > 0) |
|
|
($line.value.items | map(item_status)) as $statuses |
|
|
expected_status($statuses) as $want |
|
|
($line.value | aggregate_status) as $got |
|
|
select($want != $got) |
|
|
"line \($line.key + 1) source_epic=\($line.value.source_epic // "<missing>") aggregate=\($got) items=\($want)"
|
|
' "$LIVE_QUEUE" 2>/dev/null || true
|
|
)"
|
|
if [[ -n "$lifecycle_drift" ]]; then
|
|
fail "next-work.jsonl has aggregate/item lifecycle drift: $(printf '%s\n' "$lifecycle_drift" | head -1)"
|
|
fi
|
|
fi
|
|
|
|
if [[ "$failures" -gt 0 ]]; then
|
|
echo "next-work contract parity validation FAILED ($failures finding(s))." >&2
|
|
exit 1
|
|
fi
|
|
|
|
echo "next-work contract parity validation passed."
|