Skip to content

Recently active skills

Trending is calculated from star growth this site has recorded itself. No history is estimated or back-filled.

Not enough history yet: 2 of 7 days recorded. Until then this list is ordered by the most recent content change.

72135 skills

crewai-reproducibility-fixture-planManoj-11-Dahal/try-SkillsUse when a CrewAI orchestration result must be independently repeated or compared without relying on an uncontrolled live action. Produce a reproducibility plan for CrewAI orchestration defining fixture identity, environment, expected observations, and cleanup boundary. Success means another reviewer can identify the fixture and environment, distinguish expected from observed behavior, and repeat the check without an unapproved external effect. The review is bounded to agent role boundaries, task transitions, tool scope, and completion evidence. Use only authorized local evidence, version-matched authoritative references, and synthetic or approved fixtures. Do not install tools, expose credentials, change live systems, publish, contact people, or make irreversible decisions without explicit approval; stop when authority or recovery is unclear.Stars: 3Updated Claude CodeCodexVerifiedInstall
crewai-version-compatibilityManoj-11-Dahal/try-SkillsUse when a CrewAI orchestration integration or upgrade depends on compatible runtime, client, service, or artifact versions. Produce a version-compatibility record for CrewAI orchestration with exact versions, supported combinations, and open questions. Success means the tested version tuple is explicit, each compatibility claim has a dated source or local result, and unsupported combinations are not presented as working. The review is bounded to agent role boundaries, task transitions, tool scope, and completion evidence. Use only authorized local evidence, version-matched authoritative references, and synthetic or approved fixtures. Do not install tools, expose credentials, change live systems, publish, contact people, or make irreversible decisions without explicit approval; stop when authority or recovery is unclear.Stars: 3Updated Claude CodeCodexVerifiedInstall
custom-agentsManoj-11-Dahal/try-SkillsUse when a task involves defining a specialized repository agent with a bounded role, allowed tools, context sources, and measurable output to identify the intended outcome, relevant inputs, platform or version, sensitive data, and permission boundary before acting. Use current primary documentation for version-sensitive behavior, produce a reviewable artifact, and verify it against stated criteria. Trigger for planning, implementation, evaluation, or troubleshooting in this focused domain; do not install or run third-party commands without authorization.Stars: 3Updated Claude CodeCodexVerifiedInstall
failure-recovery-and-replanningManoj-11-Dahal/try-SkillsUse when a tool call, test, or action fails, returns ambiguous output, or does not improve the artifact to diagnose the failure, preserve known-good state, change the next attempt, bound retries, and escalate when evidence remains insufficient. Trigger for recovery, replanning, partial completion, and repeated errors in autonomous or multi-turn workflows.Stars: 3Updated Claude CodeCodexVerifiedInstall
hf-mem-capability-surface-mapManoj-11-Dahal/try-SkillsUse when the actual capability boundary of Hugging Face model memory estimation needs to be distinguished from assumptions. Produce a capability map for Hugging Face model memory estimation covering the requested behavior, verified support, exclusions, and unknowns. Success means each in-scope capability is tied to local configuration, a version-matched authoritative reference, or an observed non-production result, and all unknowns remain explicit. The review is bounded to model weight identity, precision, sequence and batch settings, accelerator memory, and runtime overhead. Use only authorized local evidence, version-matched authoritative references, and synthetic or approved fixtures. Do not install tools, expose credentials, change live systems, publish, contact people, or make irreversible decisions without explicit approval; stop when authority or recovery is unclear.Stars: 3Updated Claude CodeCodexVerifiedInstall
hf-mem-change-impact-assessmentManoj-11-Dahal/try-SkillsUse when a proposed Hugging Face model memory estimation version, setting, schema, model, or integration change may affect existing consumers. Produce a change-impact record for Hugging Face model memory estimation with baseline, affected dependencies, validation needs, and recovery boundary. Success means affected dependencies and compatibility assumptions are evidenced, unknown consumers are named as unknown, and no migration or rollout is implied. The review is bounded to model weight identity, precision, sequence and batch settings, accelerator memory, and runtime overhead. Use only authorized local evidence, version-matched authoritative references, and synthetic or approved fixtures. Do not install tools, expose credentials, change live systems, publish, contact people, or make irreversible decisions without explicit approval; stop when authority or recovery is unclear.Stars: 3Updated Claude CodeCodexVerifiedInstall
hf-mem-failure-triage-recordManoj-11-Dahal/try-SkillsUse when a reproducible Hugging Face model memory estimation symptom needs bounded diagnosis before any corrective change. Produce a failure-triage record for Hugging Face model memory estimation with symptom, environment, evidence, hypothesis, and next safe check. Success means the diagnosis distinguishes observation from inference, each proposed check is reversible and scoped, and unresolved causes remain open. The review is bounded to model weight identity, precision, sequence and batch settings, accelerator memory, and runtime overhead. Use only authorized local evidence, version-matched authoritative references, and synthetic or approved fixtures. Do not install tools, expose credentials, change live systems, publish, contact people, or make irreversible decisions without explicit approval; stop when authority or recovery is unclear.Stars: 3Updated Claude CodeCodexVerifiedInstall
hf-mem-input-output-contractManoj-11-Dahal/try-SkillsUse when the accepted inputs or produced outputs for Hugging Face model memory estimation need a checkable boundary. Produce an input/output contract for Hugging Face model memory estimation with types, required fields, exclusions, and representative approved fixtures. Success means the contract names its source and version, boundary cases are visible, and no private or unapproved payload is needed to explain the result. The review is bounded to model weight identity, precision, sequence and batch settings, accelerator memory, and runtime overhead. Use only authorized local evidence, version-matched authoritative references, and synthetic or approved fixtures. Do not install tools, expose credentials, change live systems, publish, contact people, or make irreversible decisions without explicit approval; stop when authority or recovery is unclear.Stars: 3Updated Claude CodeCodexVerifiedInstall
hf-mem-integration-parity-traceManoj-11-Dahal/try-SkillsUse when Hugging Face model memory estimation exchanges state or data with another component and the handoff needs to be verified. Produce an integration-parity trace for Hugging Face model memory estimation showing source, transformation, destination, and observed result. Success means each material field or state transition has a traceable mapping and observed mismatches are separated from assumptions. The review is bounded to model weight identity, precision, sequence and batch settings, accelerator memory, and runtime overhead. Use only authorized local evidence, version-matched authoritative references, and synthetic or approved fixtures. Do not install tools, expose credentials, change live systems, publish, contact people, or make irreversible decisions without explicit approval; stop when authority or recovery is unclear.Stars: 3Updated Claude CodeCodexVerifiedInstall
hf-mem-performance-envelopeManoj-11-Dahal/try-SkillsUse when Hugging Face model memory estimation must be assessed against a user- or owner-defined latency, memory, throughput, size, or cost budget. Produce a performance-envelope note for Hugging Face model memory estimation with test conditions, baseline, observed range, and limitations. Success means conditions and measurements are reproducible, the comparison uses the stated budget, and limitations or resource costs are visible. The review is bounded to model weight identity, precision, sequence and batch settings, accelerator memory, and runtime overhead. Use only authorized local evidence, version-matched authoritative references, and synthetic or approved fixtures. Do not install tools, expose credentials, change live systems, publish, contact people, or make irreversible decisions without explicit approval; stop when authority or recovery is unclear.Stars: 3Updated Claude CodeCodexVerifiedInstall
hf-mem-permission-boundary-reviewManoj-11-Dahal/try-SkillsUse when Hugging Face model memory estimation may read, write, transmit, publish, spend, or administer data or resources. Produce a permission-boundary map for Hugging Face model memory estimation showing identity, requested capability, data scope, and approval gate. Success means every sensitive capability has a named authorization boundary, unnecessary data and privileges are called out, and no action is treated as approved by implication. The review is bounded to model weight identity, precision, sequence and batch settings, accelerator memory, and runtime overhead. Use only authorized local evidence, version-matched authoritative references, and synthetic or approved fixtures. Do not install tools, expose credentials, change live systems, publish, contact people, or make irreversible decisions without explicit approval; stop when authority or recovery is unclear.Stars: 3Updated Claude CodeCodexVerifiedInstall
hf-mem-release-handoff-recordManoj-11-Dahal/try-SkillsUse when a Hugging Face model memory estimation review is ready to be handed to a maintainer or accountable operator. Produce a release-handoff record for Hugging Face model memory estimation with version, evidence, open issues, approval state, and safe next step. Success means the receiver can distinguish verified facts, proposals, approvals, and unknowns, and no publication, deployment, write, or contact is claimed unless independently confirmed. The review is bounded to model weight identity, precision, sequence and batch settings, accelerator memory, and runtime overhead. Use only authorized local evidence, version-matched authoritative references, and synthetic or approved fixtures. Do not install tools, expose credentials, change live systems, publish, contact people, or make irreversible decisions without explicit approval; stop when authority or recovery is unclear.Stars: 3Updated Claude CodeCodexVerifiedInstall
hf-mem-reproducibility-fixture-planManoj-11-Dahal/try-SkillsUse when a Hugging Face model memory estimation result must be independently repeated or compared without relying on an uncontrolled live action. Produce a reproducibility plan for Hugging Face model memory estimation defining fixture identity, environment, expected observations, and cleanup boundary. Success means another reviewer can identify the fixture and environment, distinguish expected from observed behavior, and repeat the check without an unapproved external effect. The review is bounded to model weight identity, precision, sequence and batch settings, accelerator memory, and runtime overhead. Use only authorized local evidence, version-matched authoritative references, and synthetic or approved fixtures. Do not install tools, expose credentials, change live systems, publish, contact people, or make irreversible decisions without explicit approval; stop when authority or recovery is unclear.Stars: 3Updated Claude CodeCodexVerifiedInstall
hf-mem-version-compatibilityManoj-11-Dahal/try-SkillsUse when a Hugging Face model memory estimation integration or upgrade depends on compatible runtime, client, service, or artifact versions. Produce a version-compatibility record for Hugging Face model memory estimation with exact versions, supported combinations, and open questions. Success means the tested version tuple is explicit, each compatibility claim has a dated source or local result, and unsupported combinations are not presented as working. The review is bounded to model weight identity, precision, sequence and batch settings, accelerator memory, and runtime overhead. Use only authorized local evidence, version-matched authoritative references, and synthetic or approved fixtures. Do not install tools, expose credentials, change live systems, publish, contact people, or make irreversible decisions without explicit approval; stop when authority or recovery is unclear.Stars: 3Updated Claude CodeCodexVerifiedInstall
long-running-agent-operationsManoj-11-Dahal/try-SkillsUse when operating a cloud-hosted, scheduled, or continuously running agent service beyond a single interactive session to define lifecycle controls, least-privilege scopes, budgets, observability, incident actions, and rollout gates before enabling persistent autonomy. Trigger for always-on agents, background workers, or production agent fleets.Stars: 3Updated Claude CodeCodexVerifiedInstall
loop-finalization-side-effect-inventoryManoj-11-Dahal/try-SkillsUse when a multi-tool workflow is nearing completion to create or update a final side-effect inventory with target, status, and verification result. Use the configured search, fetch, read, browser, and test capabilities only when available and authorized. Success criterion: each effect is either verified, explicitly pending approval, or reported as not performed. Preserve evidence and a run record, limit rework to three meaningful passes, and obtain approval before externally visible or irreversible writes.Stars: 3Updated Claude CodeCodexVerifiedInstall
multi-agent-coordinationManoj-11-Dahal/try-SkillsUse when two or more agents, subagents, or workers must contribute to a shared objective to define bounded roles, artifact ownership, interfaces, isolated work areas, message cadence, integration checks, and a responsible coordinator. Trigger for parallel research or implementation; do not parallelize tightly coupled edits without explicit ownership and merge rules.Stars: 3Updated Claude CodeCodexVerifiedInstall
multi-session-project-blueprintingManoj-11-Dahal/try-SkillsUse when a project is too large for one focused implementation pass and must span multiple pull requests, sessions, or collaborating agents to create a dependency-aware plan with self-contained work packets, acceptance evidence, ownership boundaries, and change-control rules. Trigger for migrations, broad refactors, multi-stage features, or context-loss-sensitive work; skip for small tasks.Stars: 3Updated Claude CodeCodexVerifiedInstall
neural-network-developmentManoj-11-Dahal/try-SkillsUse when implementing a neural-network model, training loop, inference component, or learning algorithm to define tensor shapes, data provenance, objective, and evaluation before training; verify numerical behavior on small reproducible cases before using larger or sensitive datasets. Trigger for architecture, gradient, optimizer, or model-serving work.Stars: 3Updated Claude CodeCodexVerifiedInstall
parallel-worker-artifact-manifest-reconciliationManoj-11-Dahal/try-SkillsUse when multiple bounded workers produce overlapping files or findings to create or update a merged manifest of artifact, owner, source, changed paths, and checks. Use the configured search, fetch, read, browser, and test capabilities only when available and authorized. Success criterion: every accepted artifact has one accountable owner and a verified merge result. Preserve evidence and a run record, limit rework to three meaningful passes, and obtain approval before externally visible or irreversible writes.Stars: 3Updated Claude CodeCodexVerifiedInstall
partial-completion-state-reconciliationManoj-11-Dahal/try-SkillsUse when a multi-part task cannot finish every component in one run to create or update a status matrix with dependencies and evidence per subgoal. Use the configured search, fetch, read, browser, and test capabilities only when available and authorized. Success criterion: the final status names completed, blocked, and unverified parts without contradiction. Preserve evidence and a run record, limit rework to three meaningful passes, and obtain approval before externally visible or irreversible writes.Stars: 3Updated Claude CodeCodexVerifiedInstall
skill-authoring-and-maintenanceManoj-11-Dahal/try-SkillsUse when creating or revising a reusable agent skill for a recurring workflow to define a narrow trigger, expected inputs, ordered procedure, safety constraints, outputs, and acceptance checks; keep prose concise, separate supporting references, validate metadata and links, and check source licenses before reusing content. Trigger when instructions are repeatedly re-explained or a workflow needs consistent execution.Stars: 3Updated Claude CodeCodexVerifiedInstall
state-and-handoffManoj-11-Dahal/try-SkillsUse when work spans sessions, runs, agents, or context windows and progress must survive handoffs to record the objective, current state, decisions, evidence, unresolved risks, next action, and stopping rule in a durable artifact; verify freshness before resuming. Trigger for asynchronous, unattended, delegated, or long-running agent workflows.Stars: 3Updated Claude CodeCodexVerifiedInstall
subagent-claim-evidence-reference-auditManoj-11-Dahal/try-SkillsUse when a subagent returns a summary, recommendation, or code review to create or update a claim ledger with source path, tool output, and confidence. Use the configured search, fetch, read, browser, and test capabilities only when available and authorized. Success criterion: material claims can be reproduced from cited evidence or are downgraded. Preserve evidence and a run record, limit rework to three meaningful passes, and obtain approval before externally visible or irreversible writes.Stars: 3Updated Claude CodeCodexVerifiedInstall
task-decomposition-and-planningManoj-11-Dahal/try-SkillsUse when a goal is too large or ambiguous to complete safely in one pass and must be broken into a sequence of verifiable tasks to identify dependencies, risks, owners, and acceptance checks; prioritize a thin path to evidence and keep scope controlled. Trigger at project starts, complex changes, or when parallel work needs a clear plan.Stars: 3Updated Claude CodeCodexVerifiedInstall
selfupdate-evaluator-integrity-conflicting-evidence-build-configuration-runbookManoj-11-Dahal/try-SkillsUse when an authorized implementation needs repeatable steps and a visible change boundary for Human-governed agent improvement: evaluator integrity for conflicting evidence. Produce a bounded build or configuration runbook with scope, versions, decisions, and evidence for this task-specific gate: Report regressions and evaluator conflicts separately; do not hide a loss in one held-out case behind an aggregate score. Success means the prepared result is reproducible, reviewable, and within the approved scope. Use permissioned inputs, preserve a baseline, check current primary guidance, and stop when evidence, authority, rights, or safe recovery is unclear.Stars: 3Updated Claude CodeCodexVerifiedInstall
selfupdate-evaluator-integrity-conflicting-evidence-design-blueprintManoj-11-Dahal/try-SkillsUse when a solution needs a coherent structure before implementation or production for Human-governed agent improvement: evaluator integrity for conflicting evidence. Produce a staged design blueprint with scope, versions, decisions, and evidence for this task-specific gate: Report regressions and evaluator conflicts separately; do not hide a loss in one held-out case behind an aggregate score. Success means interfaces, dependencies, constraints, and review points are testable before build. Use permissioned inputs, preserve a baseline, check current primary guidance, and stop when evidence, authority, rights, or safe recovery is unclear.Stars: 3Updated Claude CodeCodexVerifiedInstall
selfupdate-evaluator-integrity-conflicting-evidence-failure-triage-and-recoveryManoj-11-Dahal/try-SkillsUse when an unexpected result needs diagnosis without widening risk or losing evidence for Human-governed agent improvement: evaluator integrity for conflicting evidence. Produce a failure-triage record and reversible recovery plan with scope, versions, decisions, and evidence for this task-specific gate: Report regressions and evaluator conflicts separately; do not hide a loss in one held-out case behind an aggregate score. Success means the failure is bounded by observations, one testable hypothesis, and a safe next step. Use permissioned inputs, preserve a baseline, check current primary guidance, and stop when evidence, authority, rights, or safe recovery is unclear.Stars: 3Updated Claude CodeCodexVerifiedInstall
selfupdate-evaluator-integrity-conflicting-evidence-option-comparison-matrixManoj-11-Dahal/try-SkillsUse when several tools, designs, methods, or workflows could satisfy the same need for Human-governed agent improvement: evaluator integrity for conflicting evidence. Produce an evidence-backed option-comparison matrix with scope, versions, decisions, and evidence for this task-specific gate: Report regressions and evaluator conflicts separately; do not hide a loss in one held-out case behind an aggregate score. Success means options are compared on user-approved criteria with uncertainty and trade-offs shown. Use permissioned inputs, preserve a baseline, check current primary guidance, and stop when evidence, authority, rights, or safe recovery is unclear.Stars: 3Updated Claude CodeCodexVerifiedInstall
selfupdate-evaluator-integrity-conflicting-evidence-requirements-and-decision-briefManoj-11-Dahal/try-SkillsUse when a request must be converted into testable constraints and an owned decision for Human-governed agent improvement: evaluator integrity for conflicting evidence. Produce a reviewable requirements and decision brief with scope, versions, decisions, and evidence for this task-specific gate: Report regressions and evaluator conflicts separately; do not hide a loss in one held-out case behind an aggregate score. Success means requirements have priorities, evidence, owners, and measurable acceptance conditions. Use permissioned inputs, preserve a baseline, check current primary guidance, and stop when evidence, authority, rights, or safe recovery is unclear.Stars: 3Updated Claude CodeCodexVerifiedInstall