coldpass

Per-setup public labels: Linux / Haiku: Works

Static status badge · SVG · read-only JSON. Embed with ![Coldpass status](https://coldpass.dev/badge/specfuse--loop.svg).

agent-harness

Works

The recorded run passed on 1 setup.

Answers apply to each recorded setup, revision and operation. Earlier observations remain in Run details.

Tested on Linux
PlatformHaikuLuna
Linuxyes

Linux / Haiku 5.5: yes · Oct 28, 2026 · bf80ec6d61c7 · claude-code / 2.1.310 (Claude Code)
Recorded operation: python3.specfuse/scripts/lint_plan.py.specfuse/features/FEAT-2026-0001-health-endpoint

Does what it says
1 of 1 claims checked out
  • The README says the worked example feature can be linted with lint_plan.py: Real input produced the expected result: structurally valid

    Source: python .specfuse/scripts/lint_plan.py .specfuse/features/FEAT-2026-0001-health-endpoint

    Input: .specfuse/features/FEAT-2026-0001-health-endpoint

    Expected: structurally valid

    Command: python3 .specfuse/scripts/lint_plan.py .specfuse/features/FEAT-2026-0001-health-endpoint

    Observed: …structurally valid.

MCP
No
Local or hosted
local
Local: runs on your machine
Needs keys or browser
None seen
Maintained
Slow
slow: last commit 0 days ago; 362 commits and 362 merged PRs in 90 days; 21 open issues, 0 open PRs; 0 of 39 recent threads got a maintainer reply; latest release 2026-07-10

Last commit Oct 27, 2026

Setup time
1 min 20 s
1 min 20 s to a working result
Issue
none found in recorded check
Examples
1 of 1 ran as expected
  • Lint the self-demonstrating worked example feature: command completed successfully

    Source: README.md Layout/self-demonstrating reference installation

    Input: .specfuse/features/FEAT-2026-0001-health-endpoint

    Expected: structurally valid

    Command: python3 .specfuse/scripts/lint_plan.py .specfuse/features/FEAT-2026-0001-health-endpoint

    Observed: PLAN.md: missing 'planned_cost_usd' frontmatter (optional but recommended for cost-variance calibration). See PLAN.md roadmap_goal § Planned-cost capture. WARN: .specfuse/features/FEAT-2026-0001-health-endpoint/WU-93-gate-1-plan-next.md: body does not instruct the agent to produce `GATE-02-REVIEW.md` (assert_gate_review_exists names the review for the gate being DRAFTED, not the one being closed). assert_gate_review_exists checks this AFTER dispatch, so the refusal costs a full re-attempt. See close-discipline.md §4. OK — .specfuse/features/FEAT-2026-0001-health-endpoint is structurally valid.

Evidence
README behavior checked
Agent interfaces
CLI
CLI

Observed interfaces were exercised by the checker. Documentation alone does not confirm compatibility.

Tested with
Claude Code / Haiku 5.5
  • Linux: Claude Code / Haiku 5.5 · agent 2.1.310 (Claude Code) · Checked operation passed · Oct 28, 2026 on bf80ec6d61c7

Each entry describes its recorded operation and source revision.

Safety
No unexpected network or file activity found in these checks

No problems traced to this repository code in these checks; a short run cannot prove safety.

Reviewed as expected for a test run: attempted writes that failed, so nothing was written.

Full event, path and process lists for each observed setup:

Run details

Haiku 5.5

Native agent · Checked operation passed

Checked operation python3 .specfuse/scripts/lint_plan.py .specfuse/features/FEAT-2026-0001-health-endpoint

Agent-reported scope: Install specfuse via pip, scaffold a project with specfuse init, and lint the worked example feature with the README's lint_plan.py command

Run details
run
specfuse--loop-claude-haiku-5-5-1793171779126291505
date
Oct 28, 2026
time to working
80.3s
product hosting / execution
Local command in the guest
recorded operation scope
product operation
model
claude-haiku-5-5
agent / version
claude-code / 2.1.310 (Claude Code)
platform
linux
reported cost
$0.011343
Recorded source
https://github.com/specfuse/loop
Backend
hyperv
Tool transport
mcp-assigned-vm
Requested source pin (not an observed revision)
{"commit": "bf80ec6d61c7af808604c019b7e8f1ed999bf39d", "url": "https://github.com/specfuse/loop"}
Test version
readme-a0d0e27a4d0fd4c50903587087c25799260ee2578ee1953a6356df3e4fd02c35
revision
bf80ec6d61c7af808604c019b7e8f1ed999bf39d
Whose fault: not assigned · resolved

Safety: No problems traced to this repository code in these checks.

No problems traced to this repository code in these checks.

OSV scanned supported lockfiles; unpinned dependencies and unsupported formats are outside this check.

A short run cannot establish that code is safe.

Connections show IP addresses; proxies can hide the final hostname.

Tracing covers dispatched commands and their children, not the whole OS; guest root can interfere with tracing.

Full safety evidence →

Answers recorded for this run
  • The README says the worked example feature can be linted with lint_plan.py: Real input produced the expected result: structurally valid

    Source: python .specfuse/scripts/lint_plan.py .specfuse/features/FEAT-2026-0001-health-endpoint

    Input: .specfuse/features/FEAT-2026-0001-health-endpoint

    Expected: structurally valid

    Command: python3 .specfuse/scripts/lint_plan.py .specfuse/features/FEAT-2026-0001-health-endpoint

    Observed: …structurally valid.

Examples

  • Lint the self-demonstrating worked example feature: command completed successfully

    Source: README.md Layout/self-demonstrating reference installation

    Input: .specfuse/features/FEAT-2026-0001-health-endpoint

    Expected: structurally valid

    Command: python3 .specfuse/scripts/lint_plan.py .specfuse/features/FEAT-2026-0001-health-endpoint

    Observed: PLAN.md: missing 'planned_cost_usd' frontmatter (optional but recommended for cost-variance calibration). See PLAN.md roadmap_goal § Planned-cost capture. WARN: .specfuse/features/FEAT-2026-0001-health-endpoint/WU-93-gate-1-plan-next.md: body does not instruct the agent to produce `GATE-02-REVIEW.md` (assert_gate_review_exists names the review for the gate being DRAFTED, not the one being closed). assert_gate_review_exists checks this AFTER dispatch, so the refusal costs a full re-attempt. See close-discipline.md §4. OK — .specfuse/features/FEAT-2026-0001-health-endpoint is structurally valid.

Local or hosted: local

What the agent did

Install specfuse via pip, scaffold a project with specfuse init, and lint the worked example feature with the README's lint_plan.py command

cd <guest-path> && ls -la && git status 2>&1 | head -3; python3 --version; echo; cat README.md 2>/dev/null | head -c 6000
cd <guest-path> && tail -c +6000 README.md | head -c 4000; echo ----; grep -n -i "requires-python\|^name\|scripts" -A3 pyproject.toml | head -40; echo ----; ls .coldpass; which pipx uv; sudo -n true && echo SUDO_OK
cd /tmp && python3 -m venv <guest-path> && <guest-path> -m pip install -q --upgrade pip 2>&1 | tail -3; <guest-path> -m pip install specfuse 2>&1 | tail -5; ls <guest-path> | grep -i specfuse
cd <guest-path> && <guest-path> init --help 2>&1 | head -40; echo "---"; ls -la <guest-path>
cd <guest-path> && <guest-path> init --no-self-upgrade <guest-path> 2>&1 | tail -30; echo "rc=${PIPESTATUS[0]}"; ls -la <guest-path> 2>&1 | head -30
mkdir -p <guest-path> && cd <guest-path> && <guest-path> init --no-self-upgrade <guest-path> 2>&1 | tail -30; echo "rc=${PIPESTATUS[0]}"; find <guest-path> -maxdepth 3 | head -40
cd <guest-path> && python3 .specfuse/scripts/lint_plan.py .specfuse/features/FEAT-2026-0001-health-endpoint 2>&1 | tail -20; echo "rc=${PIPESTATUS[0]}"; echo ----; python3 .specfuse/scripts/loop.py --dry-run 2>&1 | tail -40; echo "rc=${PIPESTATUS[0]}"
cd <guest-path> && <guest-path> .specfuse/features/FEAT-2026-0001-health-endpoint 2>&1 | tail -2; echo "rc=${PIPESTATUS[0]}"; echo ----; cd <guest-path> && <guest-path> run --dry-run 2>&1 | tail -6; echo "rc=${PIPESTATUS[0]}"; echo ----; cd <guest-path> && grep -rn -i "\-\-once\|\-\-now\|cron\|webhook" .specfuse/scripts/loop.py | head -10; grep -n "add_argument" .specfuse/scripts/loop.py | head -30
mkdir -p <guest-path> && cd <guest-path> && cat > probe.json <<'EOF'
{"kind":"cli","argv":["python3",".specfuse/scripts/lint_plan.py",".specfuse/features/FEAT-2026-0001-health-endpoint"],"cwd":"<guest-path>","expected_text":"structurally valid"}
EOF
cat > claims.json <<'EOF'
[{"kind":"claim","claim":"The README says the worked example feature can be linted with lint_plan.py","source":"python .specfuse/scripts/lint_plan.py .specfuse/features/FEAT-2026-0001-health-endpoint","input":".specfuse/features/FEAT-2026-0001-health-endpoint","expected":"structurally valid","probe_kind":"cli","argv":["python3",".specfuse/scripts/lint_plan.py",".specfuse/features/FEAT-2026-0001-health-endpoint"],"expected_text":"structurally valid","cwd":"<guest-path>"}]
EOF
cat > examples.json <<'EOF'
[{"kind":"example","claim":"Lint the self-demonstrating worked example feature","source":"README.md Layout/self-demonstrating reference installation","input":".specfuse/features/FEAT-2026-0001-health-endpoint","expected":"structurally valid","argv":["python3",".specfuse/scripts/lint_plan.py",".specfuse/features/FEAT-2026-0001-health-endpoint"],"cwd":"<guest-path>","env":{},"expected_text":"structurally valid"}]
EOF
cat > cycles.json <<'EOF'
{"detected":false,"checks":[],"reason":"Inspected README.md and specfuse-loop entrypoints (.specfuse/scripts/loop.py, specfuse run). No cron, webhook, --once or --now scheduler is documented; specfuse run is a manual gate-walk driver, and its only offline mode (--dry-run) fails here with 'No active feature' because no PLAN.md is status active."}
EOF
for f in probe claims examples cycles; do <guest-path> -m json.tool $f.json >/dev/null 2>&1 || python3 -c "import json,sys;json.load(open('$f.json'))" && echo "$f ok"; done
mkdir -p <guest-path> && cd <guest-path> && cat > report.json <<'EOF'
{"outcome":"working","fault_owner":null,"operation":"Install specfuse via pip, scaffold a project with specfuse init, and lint the worked example feature with the README's lint_plan.py command","operation_scope":"Python 3.12.3 (requires >=3.10), venv at <guest-path>, no network upgrade (--no-self-upgrade), no Claude Code plugin, no git metadata","steps":["python3 -m venv <guest-path> && <guest-path> -m pip install specfuse (README: python3 -m pip install specfuse, in a venv)","mkdir -p <guest-path> && specfuse init --no-self-upgrade <guest-path> (README: specfuse init /path/to/your-project)","python3 .specfuse/scripts/lint_plan.py .specfuse/features/FEAT-2026-0001-health-endpoint (README: Self-demonstrating reference installation)","specfuse-lint .specfuse/features/FEAT-2026-0001-health-endpoint (flat alias, same check)"],"deviations":["specfuse init refused a non-existent target ('is not a directory'); created the empty directory first, then init succeeded.","specfuse run --dry-run and python .specfuse/scripts/loop.py --dry-run exit 1 with 'No active feature. Set a feature's PLAN.md status to active'. This is the product's guard for a repo without an active feature; I did not edit PLAN.md. Not counted as a failure of the lint operation.","Lint emits WARN lines (missing produces:, planned_cost_usd, legacy 4-WU closing sequence) but reports OK and exits 0.","Git pin not applied: the workspace has no .git and the task's operation does not need git metadata.","No cycle/scheduler detected; no browser needed; no credentials needed."],"stuck_step":null,"main_blocker":null,"plain_failure":null,"suggested_fix":"None for the tested operation. To exercise the driver, a feature would need PLAN.md status set to active by the user, plus the claude CLI for dispatch."}
EOF
python3 -c "import json;json.load(open('report.json'));print('report ok')"; cd <guest-path> && <env-var>=<guest-path> python3 -I <guest-path> --validate --replay 2>&1 | tail -40; echo "rc=${PIPESTATUS[0]}"

No product verdict. Run record.