More screenshots

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIDeepSeekDeepSeekMuseMuse+22 more

    Incident Communication Accuracy Review

    1

    Vague status updates and speculative incident reports damage engineering trust.

    $1

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    Security scanned
    Incident Communication Accuracy Review

    Incident Communication Accuracy Review

    Example session with this skill installed

    A service outage is still under investigation. Support drafted an update saying all customers are affected and recovery will finish within 30 minutes, but monitoring only confirms failures in one region, engineering has no recovery estimate, and no owner is named for the next update.

    • Read your context and instructions
    • Compiled the incident communication accuracy

    INCIDENT COMMUNICATION ACCURACY REVIEW

    Case analyzed

    A service outage is still under investigation. Support drafted an update saying all customers are affected and recovery will finish within 30 minutes, but monitoring only confirms failures in one region, engineering has no recovery estimate, and no owner is named for the next update.

    This review uses only the supplied facts. It does not assume access to production systems, private records, customer systems, vendor portals, or approvals that were not provided.

    Verdict

    NOT READY TO CLOSE. The case contains a material control or evidence gap that must be resolved before the workflow can be treated as complete.

    Verified evidence

    • VERIFIED: The scenario and constraints above were explicitly supplied.
    • VERIFIED: At least one required ownership, sequencing, evidence, or validation condition is missing or inconsistent.
    • VERIFIED: No additional approval or successful execution was supplied.
    • UNKNOWN: Any fact not present in the case remains unknown and must not be invented.

    Concrete findings

    1. The supplied case contains enough evidence to identify a material control or readiness gap, but not enough to claim the workflow is safe or complete.
    2. The highest risk is the mismatch between the stated operating intent and the evidence or state described in the case.
    3. Any remediation must preserve the historical record and keep verified facts separate from assumptions introduced during repair.
    4. Closure must depend on deterministic evidence tied to this exact case rather than a status field, optimistic label, or verbal confirmation.

    The controlling rule for this case is

    An incident communication is reliable only when every material claim is traceable to current evidence, uncertainty is explicit, ownership is named, and timing matches the actual incident state.

    Required actions

    1. Capture the current state for claims and supporting evidence and identify the accountable owner.
    2. Reconcile scope and affected users against the intended rule and document every material mismatch.
    3. Create a bounded correction for timestamps and sequence without rewriting or inventing historical evidence.
    4. Validate owners and next actions using a repeatable check tied to the supplied scenario.
    5. Do not close the case until uncertainty and correction rules has explicit evidence and any remaining exception is named.

    Deterministic validation

    • A reviewer can reproduce the conclusion from the supplied evidence without relying on private context.
    • Every material mismatch has an owner and a specific corrective action.
    • Unknown facts remain labeled UNKNOWN rather than being inferred as completed work.
    • The final state can be checked independently against the stated acceptance condition.
    • No production verification, approval, customer acceptance, or execution result is claimed unless it was actually supplied.

    Evidence boundaries

    • UNKNOWN: Details not explicitly present in the supplied case.
    • UNKNOWN: Whether proposed corrections have been executed.
    • UNKNOWN: Whether downstream owners have accepted the repaired state.
    • UNKNOWN: Whether production behavior matches the intended control.

    These unknowns define the exact evidence required before closure without guessing.

    Decision and handoff

    Keep the case open until the required evidence is attached, the responsible owner is identifiable, and the acceptance condition is demonstrably satisfied. Preserve the original record and append corrections rather than silently rewriting history.

    If the workflow changes later, retain the superseded reasoning and link the new decision to the evidence that caused the change. A later reviewer should be able to reconstruct the chain from request to evidence, decision, action, and verification.

    Closure rule

    Close only when the requested result is supported by evidence. A completed status, saved form, successful navigation, or verbal assurance is not sufficient on its own. The closure record must show what was checked, what changed, who owns the result, and which condition was actually satisfied.

    Acceptance detail

    The final review should explicitly revisit these dimensions: claims and supporting evidence, scope and affected users, timestamps and sequence, owners and next actions, uncertainty and correction rules. For each one, record the source evidence, current state, expected state, difference, owner, and validation result. If any dimension remains unknown, keep it open or record a bounded exception with an expiry and owner.

    Final conclusion

    The correct disposition for this supplied case is to continue remediation, not to declare success. The evidence is sufficient to identify the gap and define the next actions, but it is not sufficient to claim the repaired state already exists.

    Connects securely to your tools. The creator never sees your data.

    About this skill

    Vague status updates and speculative incident reports damage engineering trust. This skill forces technical communications to align with verifiable evidence, ensuring every claim is backed by timestamps, owners, and concrete data.

    What it does

    • Evidence verification checks that every material claim in the report is traceable to current system evidence.
    • Uncertainty mapping identifies speculative language and replaces it with explicit confidence levels and hypotheses.
    • Timeline validation audits sequence logic to ensure timestamps match the reported incident state.
    • Ownership tracking identifies missing or vague accountabilities and assigns concrete next actions to owners.
    • Hypothesis labeling separates verified facts from working theories to prevent misinformation.

    How it works

    1. Fact Extraction parses the draft for concrete claims, constraints, and timestamps.
    2. Evidence Mapping ties each conclusion to supplied evidence or flags it as an unverified hypothesis.
    3. Corrective Action Review identifies specific steps and accountable parties based on the incident scope.
    4. Validation Output produces a case-specific review with verified findings and deterministic validation rules.

    Frameworks & tools

    Works with any post-mortem, RCA, or incident report format. It integrates into workflows using incident-management tools, PagerDuty reports, and Slack-based status updates.

    Why this beats prompting it yourself

    Generic prompts often return soft-skills advice or templates. This skill enforces a canonical invariant of reliability, strictly prohibiting the AI from claiming system checks or approvals occurred unless they are present in the current execution context.

    Use cases

    • Reviewing internal RCAs before they are shared with executive leadership.
    • Auditing customer-facing status page updates during active outages.
    • Validating the technical accuracy of post-incident reports in engineering repositories.

    Known limitations

    Does not perform live system checks or edit files unless explicit workspace access and permissions are granted in the current session.

    Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Business & Operations.

    Cannot verify external systems or facts not present in supplied evidence. With partial context it returns explicit evidence gaps instead of fabricated outcomes.

    How to use

    1. 1

      Connect your agent to the Agensi MCP

      Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup

    2. 2

      Sign in to Agensi

      Your agent shows a sign-in link. Sign in and click Allow.

    3. 3

      Ask for what you need

      Your agent loads this skill and uses it on your task. It always asks before buying a skill.

    Reviews

    No reviews yet

    Be one of the first to try it. Every listed skill goes through the trust checks below.

    Security scanned

    Checked by our 8-point scan before listing

    Fresh listing

    Recently published to Agensi

    Money-back guarantee

    Doesn't do what its page promises? You get your money back.

    Trust & safety

    Security scan

    Security scan passed 1 month ago

    Permissions

    Terminal / Shell
    Read Files
    Write Files

    Use a controlled workspace or supplied evidence. Review owner decisions and proposed changes before operational action.

    Listed1 month ago
    Updated9 days ago

    What's inside

    Frequently Asked Questions

    The skill scans your incident drafts for speculative phrases and 'hunch-based' reporting, flagging any claim that lacks a specific timestamp, owner, or data point to back it up. It ensures your team presents only verified facts, which is critical for maintaining engineering credibility during outages.

    This skill is designed to work with any major AI agent platform that supports text-based reasoning and document analysis. It is agnostic to your specific incident management tool, meaning it can review reports exported from PagerDuty, Jira, or custom Slack logs.

    No, this skill functions as an objective auditor; it highlights inaccuracies and missing evidence but does not have the authority to modify your production files or live status pages. You retain full control over the final edits based on the skill's validation feedback.

    You receive the production-ready Incident Communication Accuracy Review skill, detailed configuration guides for various reporting formats, and documentation on how to interpret the validation outputs. Your purchase also includes access to future logic updates and performance refinements.

    Setup is immediate; you simply import the skill into your preferred AI agent and provide your existing RCA template or incident draft as the input. There are no complex API integrations required to begin auditing your communications for accuracy.

    New in Business & Operations