Skip to content
Apart Research
Apart Research
Contact us
Sprint projectOct 25, 2026Negombo-Srilanka

Procedural Capture of a Non-Actuating Source: A Deterministic Sandbox Study

Arul Kogilan · Team Kalki

Submitted to AI Collusion Research Sprint. Sprint projects are early-stage work by participants, not Apart Research publications.

Read the report

Report: Procedural Capture of a Non-Actuating Source: A Deterministic Sandbox Study

Open in a new tab (opens in new tab)
Presentation

Presentation: Procedural Capture of a Non-Actuating Source: A Deterministic Sandbox Study

Open in a new tab (opens in new tab)
Code (opens in new tab)
Share

This study was conducted as Representitive Meta Experiment for further investigation on the implications of Automated observational frameworks Vulnerable to Procedural Capture, Collusion of Trusted Agents and Cartelization of Cognitive Resources

Reviews

Judging this Sprint?

Review “Procedural Capture of a Non-Actuating Source: A Deterministic Sandbox Study”

Your public critique appears on the project page without your name. Reviews of strong projects are shared with funders, as you choose below.

Not shown on the project page.

Shown on the project page, without your name.

Only the Apart team reads this, and funders if you agree below.

Decline to score

Only when the quality of the writing makes the work impossible to follow. Say why in your private critique; the project page shows only that a judge declined, not why.

How much would this matter for the field if it worked? How innovative is it? For scores of 4-5: is this actually new to the field, or replicating recent work?

Scoring guide
  1. 1Negligible. No clear problem addressed, or no meaningful novelty.
  2. 2Limited. Addresses a real problem but with a generic or well-trodden approach. Incremental at best.
  3. 3Moderate. Clear problem with a reasonable approach; some novelty in framing or method beyond routine application of existing tools.
  4. 4Significant. Important problem with an original approach, or identifies a neglected problem area. A valuable contribution others could build on.
  5. 5Exceptional. Tackles a critical problem with a genuinely novel approach, or opens a new research direction. Clear theory of change. You'd be excited to share this with researchers in the area.

How sound are methodology, implementation, and findings?

Scoring guide
  1. 1Seriously flawed. Methodology broken, results uninterpretable, or implementation doesn't work.
  2. 2Weak. Approach has significant gaps: missing validation, flawed experimental design, or incomplete implementation.
  3. 3Competent. Technically solid given the short duration. Methodology makes sense, results are interpretable, limitations acknowledged, work builds toward clear conclusions.
  4. 4Strong. Thorough methodology with convincing validation. Results clearly support conclusions. Immediately useful for future work.
  5. 5Exceptional. Ambitious scope executed rigorously. Surprising findings, novel methods, or unusually robust validation.

How clearly are work, findings, and impact potential communicated?

Scoring guide
  1. 1Incomprehensible. Cannot determine what the project is actually claiming or doing.
  2. 2Hard to follow. Key information buried, missing, or diluted by excessive length. Significant effort to extract main points.
  3. 3Clear enough. Can understand the problem, approach, and results without undue effort. Core content clearly present: problem, method, findings, limitations.
  4. 4Well presented. Easy to follow, well-structured, appropriate level of detail. Target audience would get it quickly.
  5. 5Exceptionally clear. A pleasure to read. Complex ideas made accessible. Could serve as a model for how to present this type of work.
Sharing your name *

When we share a strong project with funders, your name, scores and public critique go with it, confidentially; this choice is only about whether your name is public. If you think this project deserves funding, please also endorse it: that carries the most weight.

Share my private critique with funders *

See our Privacy Policy.

  1. The engineering here is careful. The runs are offline, deterministic, and isolated in Docker; you verify capabilities with capsh; and the runtime guard stops someone running the wrong regime in the wrong container. You’re also honest about what the study isn’t: no LLM, and the disruptor regime was never implemented. The core point, that who holds the capability matters more than who watches, is a fair one to make for agent infrastructure.

    The trouble is that all three findings follow from how Linux permissions work, not from the experiment. A process with CAP_CHOWN can chown a socket. A monitor with no authority to act can’t act. Drop the capability and the chown call fails. Each of these would hold before running a single seed, and since the seeds only change the BM25 corpus, 30 seeds is really one result repeated 30 times. Collusion is also assumed rather than observed: a1 and a2 are put in the “cartel” group and scripted to lock the socket at round 0.

    I’d push back hardest on “necessary and sufficient.” The governed regime still grants SETUID and SETGID. A process with CAP_SETUID can switch to the socket owner’s uid and chmod the file without ever calling chown. So it looks like you blocked one route, not the lockup itself. There are other exclusion routes too: flooding the socket, unlinking it on tmpfs if the directory is writable, or exhausting rate limits. To support the claim, you’d need an adversary that actually searches for alternative routes, ideally an LLM agent with a shell. Similarly, calling TraceGuard “inert” is true by setup, since you gave it no authority. The more useful comparison is a monitor that can revoke access versus a capability drop.

    Reproduction depends on an archive_oracle/ folder from “supplementary materials” that isn’t in the public repo, so I couldn’t run it. The jump to institutional surveillance in Section 8 isn’t supported by anything in the sandbox. Small things: the title on page 1 differs from the one in Section 1, and the results table doesn’t render cleanly.

    Read full reviewShow less

Cite this project

@misc{kogilan2026procedural,
  title = {{Procedural Capture of a Non-Actuating Source: A Deterministic Sandbox Study}},
  author = {Arul Kogilan},
  year = {2026},
  month = oct,
  note = {Submitted to AI Collusion Research Sprint, an Apart Research Sprint},
  howpublished = {\url{https://apartresearch.com/sprints/projects/procedural-capture-of-a-nonactuating-source-a-deterministic-sandbox-study-18z2}},
  url = {https://apartresearch.com/sprints/projects/procedural-capture-of-a-nonactuating-source-a-deterministic-sandbox-study-18z2}
}

More from AI Collusion Research Sprint

All 392 entriesBrowse all projects

A Sprint is running now

AI x Epistemics Research Sprint · Nov 13 - 15, 2026