Add requesting-human-help skill for structured human-in-the-loop collaboration#673
Open
pratyush618 wants to merge 2 commits intoobra:mainfrom
Open
Add requesting-human-help skill for structured human-in-the-loop collaboration#673pratyush618 wants to merge 2 commits intoobra:mainfrom
pratyush618 wants to merge 2 commits intoobra:mainfrom
Conversation
…aboration Implements skill from issue obra#594: structured, evidence-driven requests for capability limits and high-risk actions, with validation chain and audit trail from request through human action to agent decision.
📝 WalkthroughWalkthroughAdds a new skill documentation file that defines a standardized, evidence-driven workflow and template for agent requests for human assistance, covering required fields, validation rules, audit-trail chaining, red flags, usage scenarios, and a common mistakes table with fixes. Changes
Sequence Diagram(s)sequenceDiagram
participant Agent as Agent
participant Human as Human
participant EvidenceStore as EvidenceStore
participant Auditor as Auditor
Agent->>Human: Submit structured REQUEST (goal, steps, prerequisites, acceptance criteria)
Human->>EvidenceStore: Perform action & upload EVIDENCE (screenshots, logs, outputs)
Human-->>Agent: Respond with EVIDENCE + human notes
Agent->>EvidenceStore: Validate EVIDENCE against acceptance criteria
alt Evidence meets criteria
Agent->>Auditor: Log chain: REQUEST → HUMAN ACTION → EVIDENCE → AGENT DECISION
Agent-->>Human: Confirm completion / continue workflow
else Evidence missing or fails
Agent-->>Human: Minimal follow-up request for missing pieces or escalate
end
Estimated code review effort🎯 2 (Simple) | ⏱️ ~12 minutes Poem
🚥 Pre-merge checks | ✅ 5✅ Passed checks (5 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. ✨ Finishing Touches🧪 Generate unit tests (beta)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 2
🤖 Prompt for all review comments with AI agents
Verify each finding against the current code and only fix it if needed.
Inline comments:
In `@skills/requesting-human-help/SKILL.md`:
- Around line 38-66: Add a language tag (use "text") to the triple-backticked
fenced code blocks in the "Human Help Needed" section of SKILL.md to satisfy
markdownlint MD040; locate the fenced blocks that start/stop around the "##
Human Help Needed" content (and the additional blocks noted at ranges
corresponding to the acceptance criteria and REQUEST examples) and change the
opening "```" to "```text" for each block so all fenced sections are annotated
consistently.
- Around line 34-66: Add an explicit "Human involvement level" field to the
request template so each help request records whether the human is being asked
for "Clarification", "Execution", or "Approval/Takeover": update the structured
block under "## Human Help Needed" to include a line like "**Human involvement
level:** [Clarification | Execution | Approval/Takeover]" and, where relevant,
ensure "Steps", "Expected output / evidence needed", and "Acceptance criteria"
reference or differentiate behavior based on that level (e.g., approval requires
explicit confirm/accept output while clarification may require only answers), so
reviewers can unambiguously determine the authority granted by the human
response.
ℹ️ Review info
⚙️ Run configuration
Configuration used: Repository UI
Review profile: CHILL
Plan: Pro
Run ID: 38f9aabf-63f4-498d-93b7-bac4b6b92ec0
📒 Files selected for processing (1)
skills/requesting-human-help/SKILL.md
…uage tags - Add explicit involvement level (clarification/execution/approval) to the request template so the audit trail records what authority the human response grants - Add 'text' language tag to all fenced blocks to satisfy MD040
There was a problem hiding this comment.
♻️ Duplicate comments (1)
skills/requesting-human-help/SKILL.md (1)
62-103:⚠️ Potential issue | 🟠 MajorMake evidence and validation branch on the involvement level.
approval/takeoveris now named, but the template still treats its proof the same way asexecution. For approval flows, the key artifact is the explicit authorization itself, including scope/constraints; otherwiseREQUEST → HUMAN ACTION → EVIDENCE → AGENT DECISIONstill doesn't prove what authority was granted.Suggested doc update
**Expected output / evidence needed:** -- [What to capture: screenshot, log output, command result, confirmation text] -- [Format: paste text output, attach screenshot, confirm yes/no] +- clarification: [Direct answer, plus cited source/screenshot if needed] +- execution: [Artifact proving the step ran: screenshot, log output, command result] +- approval/takeover: [Exact approval/takeover statement, scope, constraints, and who owns the action] **Acceptance criteria:** -- [ ] [Specific, verifiable condition that means "this worked"] -- [ ] [What distinguishes success from partial success] +- [ ] [Criteria appropriate to the involvement level] +- [ ] [For approval/takeover: authorization is explicit and scoped] IF all criteria met: → State: "Confirmed: [criterion 1], [criterion 2]. Proceeding." + → For approval/takeover, restate exactly what was approved and any limits before continuing. REQUEST → [structured block above] HUMAN ACTION → [what they did] EVIDENCE → [artifact they returned] -AGENT DECISION → [what you decided based on evidence] +AGENT DECISION → [what you decided based on evidence, including approval scope/limits]🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@skills/requesting-human-help/SKILL.md` around lines 62 - 103, Summary: The validation template treats approval/takeover like execution and doesn't require explicit authorization artifacts; update the "Validating the Human Response" logic to branch on involvement level and require scope/constraints evidence for approval flows. Fix: In the "Validating the Human Response" section, add branching that checks the involvement type (e.g., "approval/takeover" vs "execution") and for approval flows require an explicit authorization artifact (authorization text, scope/constraints, signer identity) as evidence; for execution flows keep the current evidence rules (logs/screenshots/outputs). Also update the guidance text and the REQUEST→HUMAN ACTION→EVIDENCE→AGENT DECISION audit chain to note that for approval flows the EVIDENCE must include the authorization scope/constraints and signer, and adjust instructions to "If any criterion unmet → Request ONLY the missing piece" to reflect the involvement-specific missing artifact handling.
🤖 Prompt for all review comments with AI agents
Verify each finding against the current code and only fix it if needed.
Duplicate comments:
In `@skills/requesting-human-help/SKILL.md`:
- Around line 62-103: Summary: The validation template treats approval/takeover
like execution and doesn't require explicit authorization artifacts; update the
"Validating the Human Response" logic to branch on involvement level and require
scope/constraints evidence for approval flows. Fix: In the "Validating the Human
Response" section, add branching that checks the involvement type (e.g.,
"approval/takeover" vs "execution") and for approval flows require an explicit
authorization artifact (authorization text, scope/constraints, signer identity)
as evidence; for execution flows keep the current evidence rules
(logs/screenshots/outputs). Also update the guidance text and the REQUEST→HUMAN
ACTION→EVIDENCE→AGENT DECISION audit chain to note that for approval flows the
EVIDENCE must include the authorization scope/constraints and signer, and adjust
instructions to "If any criterion unmet → Request ONLY the missing piece" to
reflect the involvement-specific missing artifact handling.
ℹ️ Review info
⚙️ Run configuration
Configuration used: Repository UI
Review profile: CHILL
Plan: Pro
Run ID: 0f564151-447d-4966-abaa-a17244137054
📒 Files selected for processing (1)
skills/requesting-human-help/SKILL.md
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
skills/requesting-human-help/SKILL.mdwith a structured human-in-the-loop patternCloses #594