Research Note
AI Coding Task Autonomy Matrix
Task suitability depends on evidence and consequence, not on whether an agent can generate a plausible diff.
In this article
AI Coding Task Autonomy Matrix
Decision principle
Task suitability depends on evidence and consequence, not on whether an agent can generate a plausible diff.
Dimensions
| Dimension | Lower-risk signal | Higher-risk signal |
|---|---|---|
| Clarity | Specific outcome and examples | Ambiguous product or policy decision |
| Reversibility | Isolated diff and clean rollback | Data deletion, migration, or external mutation |
| Observability | Fast deterministic tests | Delayed, noisy, or unavailable feedback |
| Testability | Known acceptance checks | Correctness depends on hidden real-world state |
| Authority | Local read and bounded write | Secrets, production, users, money, infrastructure |
| Sensitivity | Public or synthetic data | Personal, regulated, confidential, security data |
| Blast radius | One module or sandbox | Shared platform, identity, billing, security boundary |
| Expertise | Reviewer understands the domain | No qualified reviewer can validate the result |
Autonomy levels
Research or explain is read-only and produces no repository mutation.
Propose-only may create a plan or patch for review but cannot execute commands or apply changes.
Supervised execution allows bounded local edits and commands with visible checkpoints and immediate review.
Delegated execution is limited to tasks with strong isolation, permissions, tests, logs, and rollback.
Hold means the task requires a qualified human decision, stronger controls, or new authority before agent work can begin.
Scope-change rule
Reclassify the task whenever the agent discovers a migration, new dependency, external service, secret, production system, payment, user data, authorization decision, or broader shared interface.
An originally low-risk documentation change can become a higher-risk code or policy change. The original classification does not survive material scope expansion.
Decision record
Preserve the task, dimensions, autonomy level, permissions, required evidence, reviewer, stop conditions, and final acceptance decision.
Sources
Follow the evidence.
- daltonanderson.net: how i built a go app in 4 hours with cursor aidaltonanderson.net
- owasp.org: www project top 10 for large language model applicationsowasp.org
- go.dev: getting startedgo.dev
- csrc.nist.gov: finalcsrc.nist.gov
- cursor.com: auto reviewcursor.com
- daltonanderson.ghost.io: how i built a go app in 4 hours with cursor aidaltonanderson.ghost.io
- cursor.com: 1 7cursor.com
- Spotify episodeopen.spotify.com
- cursor.com: teamscursor.com
- youtu.be: n4 J1tDwreMyoutu.be
- owasp.org: www project code review guideowasp.org
- cursor.com: privacycursor.com
- cursor.com: securitycursor.com
- slsa.dev: v1.2slsa.dev
- git-scm.com: v2git-scm.com
- cursor.com: agent best practicescursor.com
- cursor.com: data usecursor.com