microsoft/cat-agent-skills is a static website that catalogs reusable instruction sets and related packages for AI agents. People use it to search, filter, rate, and download skills for Cowork, Copilot Studio, and Scout, along with Copilot plugins and Scout automations. The catalogue entries are the skills, instructions, plugins, and settings displayed by the site.
Grades each out-of-tolerance characteristic by severity and recommends a disposition (use-as-is / rework / scrap / return-to-vendor / hold-for-review) using the deterministic disposition rules engine. Use when the user says "grade the defects", "what disposition?", "how bad is it?", "can we ship it?", or after…
Drafts the nonconformance report (NCR) and inspection summary from the graded contract payload, with every determination cited to the spec and rules. Use when the user says "draft the NCR", "write up the nonconformance", "inspection summary please", or after defect-grade completes with defects.
Reads part specs, drawings, inspection plans, certificates of conformance and measurement exports for an inspection lot, and normalizes them into the mfg.quality-inspection.v1 contract inputs. Use when the user says "clear this inspection lot", "load the inspection results", "read the spec for part ", "review lot "…
Compares actual measurements against spec tolerances for an inspection lot, computes per-characteristic statistics, flags out-of-tolerance and marginal characteristics, and cross-checks certificate claims against measured reality - deterministically. Use when the user says "check tolerances", "compare against spec"…
Ranks gaps by revenue at stake with the gaprank engine and builds the action register and suggested order for the rep to confirm. Use on "what do I fix first", "rank the gaps", "build my order", after gap-score.
Retrieves the perfect-store / picture-of-success standard, planogram and promo calendar for the outlet's channel and cluster. Use on "what's the standard for this outlet", "what should this store look like", after visit-prep.
Assembles the pre-call plan for a specific outlet - profile, last visit's open actions, promotion status, recent scan performance - into the rtl.retail-execution-perfect-store.v1 contract. Use when the rep says "prep me for ", "what's the plan for today's calls", "pre-call for store ".
Drafts the return case packet - determination, recommendation, evidence list, fraud signals, customer response draft - from the contract payload. Use to close every return case: "build the case packet", "write up the return".
Determines return eligibility against the configured policy and surfaces fraud signals deterministically - the explicit Govern step of this plugin. Use on "is this returnable", "do they get a refund", "any red flags", after reason-classify.
Retrieves the applicable return, warranty and price-match policy text and the customer's configured windows for the case. Use after return-intake, or on "what does the return policy say for ".
Classifies the stated return reason against the customer's reason-code taxonomy with the deterministic reasonclassify engine. Use after return-intake in every case, or on "what reason code is this".
Structures a return request - item, order/receipt, reason as stated, condition, serials, customer history - into the rtl.returns-refund-case.v1 contract. Use when the user says "customer wants to return", "process this return", "can they get a refund", or a return case begins.
Retrieves the OSHA recordability criteria, ISO 45001 clauses, JHA/HIRA and internal EHS policy relevant to an incident and attaches them plus prior similar incidents to the mfg.safety-incident-assist.v1 contract. Use when the user says "pull the OSHA criteria", "get the JHA for this task", "have we had this incident…
Receives a safety-incident report from a form, email or voice note and pulls it into the mfg.safety-incident-assist.v1 contract inputs. Use when the user says "log this incident", "new incident report", "someone got hurt", "intake this near miss", "process this safety report", or when a Safety Incident Agent run…
Drafts the incident summary and the root-cause-analysis (RCA) template from the assembled contract payload, with every determination cited to OSHA, ISO 45001, the JHA and internal policy. Use when the user says "draft the incident summary", "write up the incident", "give me the RCA template", "prepare the OSHA…
Routes a classified incident to the right investigator, builds the owner-notification list, and starts any OSHA regulatory clock, deterministically. Use when the user says "who investigates this", "who do we notify", "route this incident", "start the OSHA clock", "who owns this", or after severity-recordability…
Classifies a safety incident's severity, near-miss status, OSHA recordability (with the OSHA 300 column) and reportability (with the 1904.39 clock), deterministically. Use when the user says "is this recordable", "classify this incident", "is this an OSHA reportable", "what severity is this", "do we have to report…
Builds a side-by-side product comparison from the PIM extract with the deterministic productcompare engine. Use when the associate asks "what's the difference between these two", "which should I recommend", or names two SKUs.
Answers "is this on promo / does this customer get the deal" using the deterministic three-part eligibility test already computed by policyresolve. Use for any promotion, discount, or deal-eligibility question.
Takes an associate's floor question with store, role and department context into the rtl.store-associate-assist.v1 contract. Use when an associate asks "can a customer return this", "do we price match", "is this on promo", "which of these two should I recommend", or any store policy, product or promotion question.
Drafts the store readiness checklist, the shift brief in store language, and the exception escalation from the contract payload. Use to close every execution run - "draft the shift brief", "write up the readiness pack".
Ranks readiness exceptions by revenue at stake with the exceptionrank engine and raises the exception list. Use when the user asks "what do we fix first", "rank the gaps", or after readiness-check.
★not rated 93▲
+4 todayA44 tokens
originalMIT
At most 3 mods per repository are shown here, and a mod shipped inside a plugin is left to that plugin's page — the rest are on their repository pages: