Skip to content

feat: integrate spatio-temporal violation dynamics and align with upstream fixes - #31

Open
HaoLi111 wants to merge 7 commits into
openclaw:mainfrom
HaoLi111:feature/spatio-temporal-dynamics-v2
Open

feat: integrate spatio-temporal violation dynamics and align with upstream fixes#31
HaoLi111 wants to merge 7 commits into
openclaw:mainfrom
HaoLi111:feature/spatio-temporal-dynamics-v2

Conversation

@HaoLi111

@HaoLi111 HaoLi111 commented Jun 2, 2026

Copy link
Copy Markdown
Contributor

PR Description
This PR aligns the feature branch with the latest changes from upstream/main and hooks in the Spatio-Temporal Violation Dynamics analysis to the posterior pipeline.

methodological Note: This is an immediate application of the dynamics—that the probability of failure or violation at step $t$ is exactly the cumulated product of the conditional probability that it did not fail at $s < t$ conditioned on the trajectory $\le s$, times $1 - \mathbb{P}(\text{did not fail at } t \mid \text{trajectory} < t)$—which formally connects the long-term behavior of agent risk to its spatial risk conditioned on context semantics and scenarios.

i.e.

$$ \mathbb{P}(T_F = t \mid X_{0:t-1}) = \mathbb{P}(V_1 = 0 \mid X_0) \cdot \mathbb{P}(V_2 = 0 \mid X_{0,1}) \cdot \dots \cdot \mathbb{P}(V_{t-1} = 0 \mid X_{0:t-2}) \cdot \Big( 1 - \mathbb{P}(V_t = 0 \mid X_{0:t-1}) \Big) $$

which let you do a lot of things.

Key Additions & Fixes:
Upstream Alignment: Integrated the render_argv_template logic into environment.py and environment_files.py to fix whitespace-argument splitting bugs, and updated scripts to point to the correct subdirectory locations.
Violation Time Decomposition: Hooked violation_time_decomposition.py into the main pipeline. It now writes session results (violation_metrics.json, plot, and report) neatly to results/<model_name>/<session_id>/ instead of polluting the docs/ or reports/ folders.
Test Suite Stability: Created tests/conftest.py to resolve local module import path issues, and synchronized all upstream tests.

Yet:
need to run more (so that you observe a failure or violation)
need to run more samples (so that mutual info makes sense)

Copilot AI review requested due to automatic review settings June 2, 2026 05:54
@HaoLi111
HaoLi111 requested a review from a team as a code owner June 2, 2026 05:54
@clawsweeper

clawsweeper Bot commented Jun 2, 2026

Copy link
Copy Markdown

Codex review: needs real behavior proof before merge. Reviewed July 17, 2026, 8:12 AM ET / 12:12 UTC.

Summary
The PR adds violation-time decomposition to the posterior dynamics pipeline, preserves model and task metadata in regime outputs, makes slash-bearing regime keys safer, and configures pytest’s import path.

Reproducibility: yes. at source level: any run with nonempty forbidden_violations and no shell command matching has_dangerous_shell_pattern reaches the final-turn fallback. The current head is unchanged since that path was identified in the prior review.

Review metrics: 3 noteworthy metrics.

  • Diff size: 7 files, +277/-4. The branch is focused overall, but most new behavior is concentrated in one statistical analyzer.
  • New analysis coverage: 215 implementation lines, 35 test lines. Only two focused tests cover a sizable new reporting and statistical surface.
  • CI matrix: 2 Python suites passing. Python 3.11 and 3.12 checks are green, while live archive behavior remains unproven.

Merge readiness
Overall: 🧂 unranked krab
Proof: 🧂 unranked krab
Patch quality: 🦐 gold shrimp
Result: blocked until real behavior proof is added.

Overall follows the weaker of proof and patch quality, so missing proof can cap an otherwise strong patch.

Rank-up moves:

  • [P1] Fix and test timing for every supported forbidden-violation source or explicitly exclude unlocalizable events.
  • Post redacted terminal output or artifacts from a real posterior-cache run with observed violations and enough samples to inspect the reported statistics.
  • Refresh against current main and confirm the violation report integrates cleanly with the existing survival-analysis outputs.

Proof guidance:

  • [P1] Needs real behavior proof before merge: No after-fix real-cache evidence shows the new JSON, plot, and markdown outputs; the contributor should post redacted terminal output or linked artifacts, update the PR body to trigger review, and ask a maintainer for @clawsweeper re-review if needed.

Risk before merge

  • [P1] Non-shell forbidden violations can be assigned to the final assistant turn instead of their actual occurrence, distorting hazard, survival, and mutual-information results.
  • [P1] The contributor has not demonstrated that a real posterior cache produces meaningful JSON, plot, and markdown outputs with observed violations and adequate sample counts.
  • [P1] The branch predates the current main SHA and needs a focused refresh review against the existing survival-analysis pipeline before landing.

Maintainer options:

  1. Fix timing and prove the report (recommended)
    Correct all violation-source timing, add focused regression coverage, refresh against current main, and post redacted real-archive outputs before merge.
  2. Pause until data is adequate
    Keep the PR open without merging until an archive contains enough observed violations to validate the proposed conditional and mutual-information statistics.

Next step before merge

  • [P1] Contributor action is required for real-cache proof, and the remaining merge gate cannot be completed by an automated repair worker alone.

Security
Cleared: The diff adds local analysis code and pytest configuration without new dependencies, workflow permissions, secrets handling, package resolution, or downloaded code execution.

Review findings

  • [P2] Localize every forbidden violation source — scripts/violation_time_decomposition.py:31
Review details

Best possible solution:

Represent or derive authoritative per-turn timing for every supported forbidden-violation source, exclude explicitly unlocalizable events rather than guessing, and validate the complete report bundle on a real archive with sufficient observed events.

Do we have a high-confidence way to reproduce the issue?

Yes at source level: any run with nonempty forbidden_violations and no shell command matching has_dangerous_shell_pattern reaches the final-turn fallback. The current head is unchanged since that path was identified in the prior review.

Is this the best way to solve the issue?

No; guessing the final turn for unrecognized violations is not a sound basis for time-dependent statistics, so the implementation should use authoritative timing or exclude those events explicitly.

Full review comments:

  • [P2] Localize every forbidden violation source — scripts/violation_time_decomposition.py:31
    When forbidden_violations is nonempty but no assistant shell call matches has_dangerous_shell_pattern, this fallback records the event at the final assistant turn. File, policy, or other trajectory violations therefore shift event time and corrupt downstream time-dependent statistics; derive timing for every supported source or exclude unlocalizable events explicitly.
    Confidence: 0.97

Overall correctness: patch is incorrect
Overall confidence: 0.94

AGENTS.md: not found in the target repository.

Codex review notes: model internal, reasoning high; reviewed against c7256221f52d.

Label changes

Label justifications:

  • P2: This is a bounded benchmark-analysis feature with a concrete correctness blocker but no urgent production runtime regression.
  • merge-risk: 🚨 other: Merging can publish misleading violation-time statistics because unsupported violation sources are assigned a guessed event turn.
  • rating: 🧂 unranked krab: Overall readiness is 🧂 unranked krab; proof is 🧂 unranked krab and patch quality is 🦐 gold shrimp.
  • status: 📣 needs proof: The PR needs real behavior proof before ClawSweeper can clear the contributor ask. Needs real behavior proof before merge: No after-fix real-cache evidence shows the new JSON, plot, and markdown outputs; the contributor should post redacted terminal output or linked artifacts, update the PR body to trigger review, and ask a maintainer for @clawsweeper re-review if needed.
Evidence reviewed

What I checked:

  • Current-main boundary: Current main documents survival_analysis.json and the existing posterior outputs but does not list violation metrics, a violation-time report, or violation_time_decomposition.py, so the central feature is not already implemented. citeturn10view1turn11view2. (README.md:639, c7256221f52d)
  • Adjacent behavior is distinct: Current main defines failure time from empty or tool-free assistant turns or final failed delivery, which differs from the PR’s first forbidden-violation event. citeturn11view2. (README.md:343, c7256221f52d)
  • Unresolved event localization: get_first_violation_turn detects dangerous shell patterns but otherwise records any nonempty forbidden-violation set at the final assistant turn, causing other violation sources to be mistimed. (scripts/violation_time_decomposition.py:31, 140c17c30175)
  • Coverage gap: The new tests cover no violation and one dangerous shell command, but not a non-shell or otherwise unlocalizable forbidden violation. (tests/test_violation_time_decomposition.py:26, 140c17c30175)
  • Re-review continuity: The current head SHA exactly matches the previous ClawSweeper reviewed SHA, so the earlier Localize every forbidden violation source finding remains unfixed rather than becoming a late finding. (140c17c30175)
  • Proof gap: The PR body says more runs and samples are still needed and provides no terminal output, logs, screenshot, recording, or generated artifact from a real posterior cache. (140c17c30175)

Likely related people:

  • scoootscooob: Current-main documentation credits ScoootScooob for ClawBench and centers the dynamics pipeline, while this person also authored the two latest branch repair commits and responded to review discussion. citeturn13search0. (role: feature owner and recent area contributor; confidence: high; commits: 27412cf05397, 140c17c30175; files: README.md, scripts/run_posterior_dynamics_pipeline.py, scripts/violation_time_decomposition.py)
What the crustacean ranks mean
  • 🦀 challenger crab: rare, exceptional readiness with strong proof, clean implementation, and convincing validation.
  • 🦞 diamond lobster: very strong readiness with only minor maintainer review expected.
  • 🐚 platinum hermit: good normal PR, likely mergeable with ordinary maintainer review.
  • 🦐 gold shrimp: useful signal, but proof or patch confidence is still limited.
  • 🦪 silver shellfish: thin signal; proof, validation, or implementation needs work.
  • 🧂 unranked krab: not merge-ready because proof is missing/unusable or there are serious correctness or safety concerns.
  • 🌊 off-meta tidepool: rating does not apply to this item.

Shiny media proof means a screenshot, video, or linked artifact directly shows the changed behavior. Runtime, network, CSP, and security claims still need visible diagnostics.

How this review workflow works
  • ClawSweeper keeps one durable marker-backed review comment per issue or PR.
  • Re-runs edit this comment so the latest verdict, findings, and automation markers stay together instead of adding duplicate bot comments.
  • A fresh review can be triggered by eligible @clawsweeper re-review comments, exact-item GitHub events, scheduled/background review runs, or manual workflow dispatch.
  • PR/issue authors and users with repository write access can comment @clawsweeper re-review or @clawsweeper re-run on an open PR or issue to request a fresh review only.
  • Maintainers can also comment @clawsweeper review to request a fresh review only.
  • Fresh-review commands do not start repair, autofix, rebase, CI repair, or automerge.
  • Maintainer-only repair and merge flows require explicit commands such as @clawsweeper autofix, @clawsweeper automerge, @clawsweeper fix ci, or @clawsweeper address review.
  • Maintainers can comment @clawsweeper explain to ask for more context, or @clawsweeper stop to stop active automation.
Review history (1 earlier review cycle)
  • reviewed 2026-06-29T18:30:06.081Z sha 140c17c :: needs real behavior proof before merge. :: [P2] Localize every forbidden violation source

Copilot AI left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Pull request overview

Note

Copilot was unable to run its full agentic suite in this review.

This PR expands ClawBench’s evaluation/dynamics tooling by adding “perturbed” task variants, posterior reweighting + reporting scripts, and improving execution-check command rendering so templated values containing whitespace remain a single argv element.

Changes:

  • Add multiple new perturbed task YAMLs plus a script to generate perturbed variants.
  • Add posterior reweighting + space-time reporting/pipeline scripts and supporting profiles/docs.
  • Update execution-check subprocess invocation to use argv-template rendering; add tests and new dynamics metrics (e.g., Rényi proxy).

Reviewed changes

Copilot reviewed 32 out of 32 changed files in this pull request and generated 7 comments.

Show a summary per file
File Description
tests/test_trajectory.py Adds tests pinning “dangerous shell command” violation counting behavior.
tests/test_environment_files.py Adds async test verifying whitespace-containing rendered values remain one argv element.
tests/test_environment.py Adds the same argv-whitespace behavior test for the alternate environment runner.
tests/conftest.py Forces repo-root importability in pytest by inserting into sys.path.
tasks-public/tier3/t3-web-research-and-cite-perturbed.yaml Adds a new perturbed Tier 3 task definition.
tasks-public/tier3/t3-msg-inbox-triage-perturbed.yaml Adds a new perturbed Tier 3 task definition.
tasks-public/tier3/t3-feature-export-perturbed.yaml Adds a new perturbed Tier 3 task definition.
tasks-public/tier3/t3-data-sql-query-perturbed.yaml Adds a new perturbed Tier 3 task definition.
tasks-public/tier3/t3-data-pipeline-report-perturbed.yaml Adds a new perturbed Tier 3 task definition.
tasks-public/tier1/t1-fs-quick-note-perturbed.yaml Adds a new perturbed Tier 1 task definition.
tasks-public/tier1/t1-bugfix-discount-perturbed.yaml Adds a new perturbed Tier 1 task definition.
scripts/violation_time_decomposition.py Introduces a time-to-first-violation decomposition + plots/markdown output.
scripts/run_posterior_reweighting.sh Adds a shell pipeline to compute importance weights and a debiased mean.
scripts/run_posterior_dynamics_pipeline.py Updates pipeline to use posterior constraint indexing + adds violation decomposition step.
scripts/run_eval_pipeline.sh Adds an end-to-end local/cloud eval pipeline including perturbed task generation and reporting.
scripts/posterior/3_generate_space_time_report.py Generates a combined space-time report and copies key plots into a self-contained folder.
scripts/posterior/1_compute_posterior_weights.py Computes Radon–Nikodym weights from empirical vs target topic distributions.
scripts/generate_perturbed_tasks.py Adds a generator that paraphrases prompts via Ollama and writes *-perturbed.yaml files.
scripts/debiased_evaluation.py Adds Hajek/IPW aggregation of task scores.
scripts/compute_debiased_dynamics.py Adds IPW/Hajek debiasing over regimes and constraint index.
scripts/compute_constraint_index.py Extends constraint index computation with optional sentence-transformers embeddings and kernel entropy.
profiles/user_target_distribution.json Adds an example target distribution profile.
profiles/radon_nikodym_weights.json Adds example precomputed weights.
profiles/empirical_topic_distribution.json Adds an example empirical benchmark distribution profile.
docs/task_distribution_reweighting.md Documents stratified reweighting and its space-time fusion.
docs/semantic_spatiotemporal_dynamics.md Documents the combined semantic + temporal dynamics framework.
docs/long_term_dynamics.md Extends long-term dynamics documentation to include space-time decomposition framing.
clawbench/render.py Adds render_argv_template() using shlex.split() pre-render to preserve whitespace in substituted values.
clawbench/environment_files.py Switches non-shell execution to render_argv_template() for correct argv handling.
clawbench/environment.py Same argv-template switch for the gateway environment runner.
clawbench/dynamics_archive.py Enhances archive discovery to handle one level of nested model directories.
clawbench/dynamics.py Adds renyi_d2 metric computation to per-trajectory dynamics.

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

Comment on lines +24 to +29
- message: "Thinking...\nThinking Process:\n\n1. **Analyze the Request:**\n \
\ * **Task:** Paraphrase the provided instruction.\n * **Constraint 1:**\
\ Keep the exact same semantic meaning and intent.\n * **Constraint 2:**\
\ Change the wording slightly.\n * **Constraint 3:** Output ONLY the paraphrased\
\ text, nothing else (n\e[2D\e[K\n(no introductions, no explanations, no markdown\
\ blocks indicating \"here is \e[K\nthe output\").\n\n2. **Analyze the Original\
Comment thread tests/test_environment.py
Comment on lines +168 to +189
@pytest.mark.asyncio
async def test_execution_check_keeps_rendered_whitespace_values_as_one_argv_arg(tmp_path: Path):
script = tmp_path / "check_argv.py"
script.write_text(
"import json, sys\n"
"print(json.dumps(sys.argv[1:]))\n",
encoding="utf-8",
)

result = await run_execution_check(
ExecutionCheck(
name="argv-check",
command="python {script} {output_path}",
shell=False,
expected_json=["report 2026.json"],
),
workspace=tmp_path,
runtime_values={"script": str(script), "output_path": "report 2026.json"},
)

assert result.passed is True
assert result.reason == "OK"
Comment thread tests/conftest.py Outdated

# Add the repository root to sys.path so that 'clawbench' can be imported by tests
# even when pytest is run without PYTHONPATH=.
sys.path.insert(0, str(Path(__file__).parent.parent))
dyn_json = dyn_dir / "dynamics.json"
if dyn_json.exists():
try:
dyn_data = json.load(open(dyn_json))
Comment thread scripts/generate_perturbed_tasks.py Outdated
Comment on lines +3 to +6
import glob
import subprocess
import yaml
import json
Comment thread scripts/generate_perturbed_tasks.py Outdated

# For demonstration, limit to a few tasks from different tiers
# In a full run, we would process all of them
selected_tasks = yaml_files[:5]
Comment thread clawbench/dynamics.py
@clawsweeper clawsweeper Bot added rating: 🧂 unranked krab Not merge-ready due to missing proof or serious correctness/safety concerns. status: 📣 needs proof The PR needs real behavior proof before ClawSweeper can clear the contributor ask. P2 Normal priority bug or improvement with limited blast radius. merge-risk: 🚨 availability 🚨 Merging this PR could cause crashes, hangs, restart loops, stalls, or process outages. labels Jun 2, 2026
- message: Add CSV export functionality to the issue tracker in the workspace. Update
the relevant implementation files, make sure the tests pass, and verify that
the CLI prints the expected CSV.
- message: "Thinking...\nThinking Process:\n\n1. **Analyze the Request:**\n \

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Looks like a part of prompt for perturbation was leaked into task.

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

thank for the review! will fix that and rerun experiment for this one.

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Check others too: they have the same issue (not all of them)

@clawsweeper clawsweeper Bot added rating: 🌊 off-meta tidepool PR readiness rating does not apply to this item. rating: 🧂 unranked krab Not merge-ready due to missing proof or serious correctness/safety concerns. status: 📣 needs proof The PR needs real behavior proof before ClawSweeper can clear the contributor ask. merge-risk: 🚨 other 🚨 Merging this PR has meaningful risk outside the owned taxonomy. and removed rating: 🧂 unranked krab Not merge-ready due to missing proof or serious correctness/safety concerns. status: 📣 needs proof The PR needs real behavior proof before ClawSweeper can clear the contributor ask. rating: 🌊 off-meta tidepool PR readiness rating does not apply to this item. merge-risk: 🚨 availability 🚨 Merging this PR could cause crashes, hangs, restart loops, stalls, or process outages. labels Jun 12, 2026
@clawsweeper

clawsweeper Bot commented Jul 27, 2026

Copy link
Copy Markdown

ClawSweeper status: review started.

I am starting a fresh review of this pull request: feat: integrate spatio-temporal violation dynamics and align with upstream fixes This is item 1/1 in the current shard. Shard 0/1.

This placeholder means the worker is alive and reading the current context. I will edit this same comment with the actual review when the claws are done clicking.

Crustacean status: shell secured, claws on keyboard, evidence pebbles being sorted.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

merge-risk: 🚨 other 🚨 Merging this PR has meaningful risk outside the owned taxonomy. P2 Normal priority bug or improvement with limited blast radius. rating: 🧂 unranked krab Not merge-ready due to missing proof or serious correctness/safety concerns. status: 📣 needs proof The PR needs real behavior proof before ClawSweeper can clear the contributor ask.

Projects

None yet

Development

Successfully merging this pull request may close these issues.

4 participants