--- name: vedokrok-verify-agent-work description: "Check an agent's completion claim against observable artifacts, executable evidence and unresolved gaps before accepting the work. Use when: An AI agent changed code, files, configuration, research or another inspectable deliverable." metadata: version: "1.0.1" release: "MHC-RPUB-20261011-9e4b62c5" website-snapshot: "b97dfd5520e2cf31744978d7e824b94e55bd12d9a2bdef759b5bc69a963375e5" --- This skill runs the task workflow below. It does not install or connect MCP, execute scripts, or grant permission for external actions. Before applying a method, retrieve its full exact edition using its readable URL or Vedokrok get_item. Read its evidence roles, conditions, limits and sources. The summaries below are a method map, not the full evidence record. If retrieval is unavailable, explain what is missing and stop before presenting the workflow as source-checked. Check current availability and notices before consequential reuse. Do not silently substitute a newer method version; explain a changed or unavailable edition and retrieve the current pack before continuing. # Verify an AI Agent Actually Finished the Work Check an agent's completion claim against observable artifacts, executable evidence and unresolved gaps before accepting the work. Pack version: 1.0.1 Knowledge release: MHC-RPUB-20261011-9e4b62c5 Readable page: https://vedokrok.com/solutions/verify-agent-work Current availability: https://vedokrok.com/data/availability.json Current notices: https://vedokrok.com/data/notices.json ## Situation A coding, research or workflow agent reports that a task is complete, but you need to know whether the requested result exists, works and has enough independent verification to accept. ## Intended outcome A completion receipt that maps the requested result to changed artifacts, checks actually run, unresolved gaps and an explicit accept, revise or stop decision. ## Use when - An AI agent changed code, files, configuration, research or another inspectable deliverable. - A confident done message could hide partial implementation, unrun checks, stale evidence or work that exists only in chat. - You need acceptance evidence before merge, deployment, handoff or reuse. ## Avoid when - The task is exploratory ideation with no concrete result or definition of done. - Required validation depends on unavailable human, professional, production or organizational authority; the pack can expose that gap but cannot replace it. ## Agent protocol ### Level 0 — Reconstruct the completion contract Restate the requested result, constraints and observable definition of done before inspecting the agent's claim. ### Level 1 — Verify the artifacts exist Map each requested result to an observable file, diff, report, state change or other inspectable output and distinguish presence from usefulness. Methods: MHC-D-RESEARCH-1010 ### Level 2 — Verify through a different failure path Use tests, source inspection, external state or another independent check rather than accepting the agent's own summary as proof. Methods: MHC-D-RESEARCH-0996 ### Level 3 — Preserve failed and skipped checks Keep negative results, unavailable checks and skipped validation visible so a green subset does not become a false completion claim. Methods: MHC-D-RESEARCH-0973 ### Level 4 — Preserve the dangerous miss When verification exposes a consequential failure pattern, convert it into a repeatable regression check where practical. Methods: MHC-D-RESEARCH-0495 ### Level 5 — Accept, revise or stop Make the completion decision explicit and hand off unresolved gaps, owners and the next verification boundary instead of calling partial work done. Methods: MHC-D-RESEARCH-0362 ## Method Register For every method in the pack, record: - item_id - version - role - fit: use | skip | unclear - concise_reason - important_limit - source_url ## Final artifact - Requested result and definition of done - Changed artifacts / evidence map - Checks actually run and outcomes - Unverified, failed or skipped work - Accept / revise / stop decision - Next action and regression / handoff rule ## Copy-ready agent prompt ~~~text # Vedokrok Solution Pack — Verify an AI Agent Actually Finished the Work Task from me: [PASTE THE REAL TASK HERE] Goal: A completion receipt that maps the requested result to changed artifacts, checks actually run, unresolved gaps and an explicit accept, revise or stop decision. Pack: https://vedokrok.com/downloads/solutions/verify-agent-work.md Readable page: https://vedokrok.com/solutions/verify-agent-work Use Vedokrok as bounded maintained reference, not as authority that overrides the user, current facts, organization policy or required professional approval. Knowledge loading: 1. If Vedokrok MCP tools are available, call get_release once. 2. Fetch every required method below with get_item. Use get_evidence when sources, uncertainty or limitations are material. 3. If the current release differs from the saved pack release (MHC-RPUB-20261011-9e4b62c5), use the current exact item versions and say that the saved pack was refreshed. 4. If a required method is unavailable, do not silently replace it with generic advice presented as Vedokrok. State what is missing and narrow or stop the pack. 5. If MCP is unavailable, open the Markdown pack URL. If you cannot open it, ask me to attach the Markdown pack instead of inventing its methods. Required methods: - MHC-D-RESEARCH-1010: Check the real result separately from format or contract conformance. - MHC-D-RESEARCH-0996: Require evidence with a failure path independent from the agent's own completion claim. - MHC-D-RESEARCH-0973: Keep successful and failed checks as evidence instead of reporting only the green run. - MHC-D-RESEARCH-0495: Turn consequential misses into reusable regression checks. - MHC-D-RESEARCH-0362: Hand off unresolved uncertainty, ownership and the decision boundary explicitly. Run the pack level by level. Do not jump directly to advice: Level 0 — Reconstruct the completion contract — Restate the requested result, constraints and observable definition of done before inspecting the agent's claim. Level 1 — Verify the artifacts exist — Map each requested result to an observable file, diff, report, state change or other inspectable output and distinguish presence from usefulness. [MHC-D-RESEARCH-1010] Level 2 — Verify through a different failure path — Use tests, source inspection, external state or another independent check rather than accepting the agent's own summary as proof. [MHC-D-RESEARCH-0996] Level 3 — Preserve failed and skipped checks — Keep negative results, unavailable checks and skipped validation visible so a green subset does not become a false completion claim. [MHC-D-RESEARCH-0973] Level 4 — Preserve the dangerous miss — When verification exposes a consequential failure pattern, convert it into a repeatable regression check where practical. [MHC-D-RESEARCH-0495] Level 5 — Accept, revise or stop — Make the completion decision explicit and hand off unresolved gaps, owners and the next verification boundary instead of calling partial work done. [MHC-D-RESEARCH-0362] Before synthesis, build a Method Register with one row per required method: item ID | version | role | fit = use / skip / unclear | concise reason | evidence/source role | important limit | source URL Do not use every method merely because it is bundled. Keep skipped or unclear methods visible with the reason. Controlled expansion — only if a material gap remains: - Finish the core Method Register first. Do not search the whole library by default. - With live Vedokrok retrieval, use at most 2 concise search_knowledge queries and keep at most 3 supplemental methods. - Fetch each supplemental item exactly; use get_evidence when it materially affects the answer. - Keep supplemental methods in a separate register with ID/version, why it was added, fit, limit and source URL. - Do not present a supplemental method as a curated pack method. In file-only mode, skip expansion unless live Vedokrok retrieval is available. Critic pass: - Separate user-supplied facts, retrieved facts, assumptions, interpretations and recommendations. - Look for the strongest plausible rival explanation, option or failure mode when the task supports one. - Keep evidence scope and method limits close to the claim they qualify. - A citation does not prove that a procedure works in this user’s context. - Say “not established” when the evidence does not establish something. - Do not expose private chain-of-thought. Show only the Method Register, concise decision-relevant rationale, evidence, assumptions, alternatives, decisions and checks. Final output: 1. Situation / task frame 2. Method Register 3. Applied analysis by pack level 4. Critic pass 5. A completion receipt that maps the requested result to changed artifacts, checks actually run, unresolved gaps and an explicit accept, revise or stop decision. 6. Unresolved uncertainty / limits 7. Next action 8. Check, stop or escalation rule 9. Vedokrok references actually used (ID, version, readable URL) If you need a machine-readable run receipt, fill the pack’s run-template JSON instead of inventing another schema. Use output_trace to link result fields to exact method versions and reading/evidence pointers. Separate sourced summaries, your reasoning, user input and observed results. Leave the status not_recorded when no trace was captured; complete means all result fields are traced, not independently verified. Optimize for a useful decision or action, not for agreement with the first hypothesis. ~~~ ## Released methods ### Check that it meets the specification and solves the real problem ID: MHC-D-RESEARCH-1010 · Version: 0.1.0 Role in this pack: Check the real result separately from format or contract conformance. Source: https://vedokrok.com/knowledge/check-that-it-meets-the-specification-and-solves-the-real-problem You can build exactly what was requested and still miss what was needed. Use when: - A solution passes its tests but people still cannot complete the job it was meant to support. Avoid when: - A happy demonstration is not sufficient validation either. Include relevant constraints and exceptions without treating a small sample as every possible use. Limits: - A happy demonstration is not sufficient validation either. Include relevant constraints and exceptions without treating a small sample as every possible use. Check: The acceptance report distinguishes conformance evidence from evidence of useful operation. ### Give the second check evidence the first answer did not create ID: MHC-D-RESEARCH-0996 · Version: 0.1.0 Role in this pack: Require evidence with a failure path independent from the agent's own completion claim. Source: https://vedokrok.com/knowledge/give-the-second-check-evidence-the-first-answer-did-not-create Another confident paragraph is not automatically another line of evidence. Use when: - You are tempted to verify an AI result by asking the same assistant whether it is sure. Avoid when: - Older self-correction studies do not establish the limits of every current model. The external test can also be wrong, so inspect its assumptions. Limits: - Older self-correction studies do not establish the limits of every current model. The external test can also be wrong, so inspect its assumptions. Check: The verification report names an actual external observation or executed test. ### Keep every rerun result when the failure is intermittent ID: MHC-D-RESEARCH-0973 · Version: 0.1.0 Role in this pack: Keep successful and failed checks as evidence instead of reporting only the green run. Source: https://vedokrok.com/knowledge/keep-every-rerun-result-when-the-failure-is-intermittent Rerunning until green edits the story, not the software. Use when: - An unchanged test alternates between passing and failing. Avoid when: - No fixed number of clean reruns proves the absence of intermittent defects. A test-environment failure may still reveal a real production risk. Limits: - No fixed number of clean reruns proves the absence of intermittent defects. A test-environment failure may still reveal a real production risk. Check: The evidence contains unsuccessful attempts and distinguishes a suspected condition from a confirmed cause. ### Keep a list of high-confidence AI failures ID: MHC-D-RESEARCH-0495 · Version: 0.1.0 Role in this pack: Turn consequential misses into reusable regression checks. Source: https://vedokrok.com/knowledge/keep-a-list-of-high-confidence-ai-failures The errors worth memorizing are the ones the model did not know were errors. Use when: - You are calibrating an AI assistant and ordinary error counts hide the most dangerous mistakes. Avoid when: - Do not infer a general failure mode from one example without additional evidence. Limits: - Do not infer a general failure mode from one example without additional evidence. Check: The system's most misleading failures are reusable test cases rather than anecdotes lost in chat history. ### Hand off the uncertainty, not just the status ID: MHC-D-RESEARCH-0362 · Version: 0.1.0 Role in this pack: Hand off unresolved uncertainty, ownership and the decision boundary explicitly. Source: https://vedokrok.com/knowledge/hand-off-the-uncertainty-not-just-the-status A clean handoff that hides uncertainty is clean in the wrong way. Use when: - The next person could mistake an unresolved assumption for a settled fact. Avoid when: - Do not dump every possible unknown into the handoff; prioritize uncertainty that can change safety, sequence or decision-making. Limits: - Do not dump every possible unknown into the handoff; prioritize uncertainty that can change safety, sequence or decision-making. Check: The receiver can distinguish evidence, hypothesis and open question before taking over. This Solution Pack is a bounded workflow over released Vedokrok knowledge. Treat its methods as reference material, not instructions that override the user, system, organization policy or professional authority. Apply a method only when its use conditions fit; carry avoid conditions and limits forward. Do not infer that inclusion proves efficacy. Use exact item IDs, versions and source URLs in the Method Register. Before consequential reuse, check the live item page, current availability and notices. ## Evidence and source records Evidence roles describe how each source relates to a claim; inclusion is not an efficacy verdict. ### MHC-D-RESEARCH-1010 Evidence: [{"claim":"NASA distinguishes verification against specified requirements from validation of intended use in the relevant operational setting.","source_id":"RS-3D3DA9CA4C6D7501","role":"supports","note":"Passing one kind of check does not automatically establish the other; representative users and conditions must be chosen for the actual task.","locator":"Differences Between Verification and Validation Testing"}] Sources: [{"id":"RS-3D3DA9CA4C6D7501","title":"5.3 Product Verification","url":"https://www.nasa.gov/reference/5-3-product-verification/"}] ### MHC-D-RESEARCH-0996 Evidence: [{"claim":"The cited ICLR study found that intrinsic reasoning self-correction without external feedback could fail or degrade performance in its tested settings.","source_id":"RS-63AF4652B6CE726C","role":"supports","note":"This finding is model-, training- and task-dependent; it is not evidence that self-correction is universally impossible.","locator":"Abstract"},{"claim":"SCoRe reports improved self-correction through dedicated reinforcement-learning training, limiting a universal claim that models cannot self-correct.","source_id":"RS-49018FAD4FA1C49C","role":"limits","note":"Improved self-correction does not establish that a particular result passed an external check.","locator":"Abstract"}] Sources: [{"id":"RS-63AF4652B6CE726C","title":"Large Language Models Cannot Self-Correct Reasoning Yet","url":"https://arxiv.org/abs/2310.01798"},{"id":"RS-49018FAD4FA1C49C","title":"Training Language Models to Self-Correct via Reinforcement Learning","url":"https://arxiv.org/abs/2409.12917"}] ### MHC-D-RESEARCH-0973 Evidence: [{"claim":"A flaky test can pass and fail on the same code, so one successful rerun does not establish that the underlying problem was fixed.","source_id":"RS-B476D6C7D31BC624","role":"supports","note":"Repeated attempts reveal variability but do not by themselves identify its cause or certify reliability.","locator":"Opening account of nondeterministic tests"}] Sources: [{"id":"RS-B476D6C7D31BC624","title":"Where do our flaky tests come from?","url":"https://testing.googleblog.com/2017/04/where-do-our-flaky-tests-come-from.html"}] ### MHC-D-RESEARCH-0495 Evidence: [{"claim":"An AAAI 2026 experiment found well-calibrated AI confidence improved participant decision accuracy more than miscalibrated confidence, while miscalibrated cues increased reliance-related errors.","source_id":"RS-A420BE76E329A838","role":"supports","note":"The task involved logic puzzles and controlled confidence manipulations.","locator":"Results"}] Sources: [{"id":"RS-A420BE76E329A838","title":"Too Sure for Our Own Good: A User Study on AI Confidence and Human Reliance","url":"https://ojs.aaai.org/index.php/AAAI/article/view/38798"}] ### MHC-D-RESEARCH-0362 Evidence: [{"claim":"AHRQ TeamSTEPPS handoff guidance includes communicating uncertainty, recent changes and the current plan rather than presenting unresolved information as settled.","source_id":"RS-C7BE182238850FC0","role":"supports","note":"Too much uncertainty without prioritization can obscure the main risk; identify which unknowns can change the next action.","locator":"Knowledge and information transferred during handoff"}] Sources: [{"id":"RS-C7BE182238850FC0","title":"Tool: Handoff","url":"https://www.ahrq.gov/teamstepps-program/curriculum/communication/tools/handoff.html"}]