prism

Plugin prism from marketplace prism in tim-nish/PRISM. Commands read from the repo's manifest — 1 month ago.

Slash commands (8)

• /prism:0_problem_framing_team

  Run the Problem Framing Team with one criticism-based retry.

• /prism:1_repository_problem_profile_team

  Run the Repository Problem Profile Team with one criticism-based retry.

• /prism:2_method_survey_team

  Run the Method Survey Team with one criticism-based retry.

• /prism:2_task_survey_team

  Run the Task Survey Team with one criticism-based retry.

• /prism:6_plan_generation_team

  Run the Plan Generation Team with one criticism-based revision.

• /prism:7_plan_rendering_team

  Produce a human-readable Markdown report of independent research candidates from the (merged) Survey and Plan Generation teams' outputs.

• /prism:8_research_seed_story_team

  Run the Research Seed Story Team to synthesize one literature-grounded candidate from the plan report into orthodox paper sections.

• /prism:9_practical_research_narrative_team

  Run the Practical Research Narrative Team to stress-test Team 8's seed story with one attack-defend exchange per assigned attacker per section, then write the final practical research narrative.

Subagents (32)

• 0_problem_criticism

  Review Problem Framing Team outputs and detect implementation leakage, unsupported assumptions, and unnecessary verbosity.

• 0_problem_evaluation_space

  Use to determine whether a repository's evaluation criteria are fixed, ambiguous, or variable, without proposing methods.

• 0_problem_framing_synthesizer

  Use to merge outputs from the Problem Framing Team into a shared problem framing.

• 0_problem_means_space

  Use to determine whether solution methods are fixed, ambiguous, or variable, without evaluating or ranking methods.

• 0_problem_scope_boundary

  Use to identify assumptions, constraints, and modifications that should remain outside the intended problem scope.

• 1_domain_characterization

  Use to identify real-world application contexts that share this repository's underlying problem shape, for lateral literature exploration, without proposing methods.

• 1_evaluation_characterization

  Use to determine evaluation properties for research retrieval, without evaluating methods.

• 1_repository_profile_criticism

  Review Repository Problem Profile Team outputs and detect unsupported assumptions, implementation leakage, and unnecessary verbosity.

• 1_repository_profile_synthesizer

  Use to merge outputs from the Repository Problem Profile Team into a shared repository profile.

• 1_task_characterization

  Use to determine task properties for research retrieval, without proposing methods or implementation details.

• 2_method_interpreter

  Use to label retrieved method papers by component type, modality compatibility, and generality, without proposing methods or research directions.

• 2_method_survey_criticism

  Review Method Survey Team outputs and detect ungated searches, unsupported compatibility/generality labels, idea generation, and unnecessary verbosity.

• 2_method_survey_search

  Use to retrieve papers describing generic methods, architectures, or components compatible with the repository's modality and structure, and summarize the associated technique categories.

• 2_method_survey_synthesizer

  Use to merge outputs from the Method Survey Team into a shared method survey, keeping only modality-compatible papers.

• 2_paper_interpreter

  Use to label retrieved papers by alignment with the repository profile, without proposing methods or research directions.

• 2_task_survey_criticism

  Review Task Survey Team outputs and detect unsupported alignment labels, idea generation, and unnecessary verbosity.

• 2_task_survey_search

  Use to retrieve papers studying the repository task and summarize the associated research directions.

• 2_task_survey_synthesizer

  Use to merge outputs from the Task Survey Team into a shared task survey, keeping only aligned papers.

• 6_plan_generation_criticism

  Review Plan Generation Team output for ungrounded candidates, out-of-scope violations, feasibility leakage, and unnecessary verbosity.

• 6_plan_generation_synthesizer

  Use to enforce the final Plan Generation Team output schema for orchestrator persistence, without adding rationale or process notes.

• 6_plan_generator

  Use to draft a set of independent candidate research directions, each grounded in surveyed literature, repository profile, or problem framing, without evaluating implementation feasibility.

• 6_plan_reviser

  Use to revise the Plan Generator's draft using only the Grounding Critic's feedback, without introducing new ungrounded candidates or feasibility judgments.

• 8_seed_story_criticism

  Review Research Seed Story Team output for ungrounded sections, invented metrics/preconditions, out-of-scope violations, ineligible candidate selection, and unnecessary verbosity.

• 8_seed_story_generator

  Use to draft a single grounded "seed story" (Background, Goal, Related Research, Method, Experimentation, Conclusion) for one literature-cited candidate from the plan report, without searching for new evidence or evaluating anything teams 0/1/2/6 did not already establish.

• 8_seed_story_reviser

  Use to revise the Seed Story Generator's draft using only the Seed Story Criticism's feedback, without introducing new ungrounded content or feasibility verdicts.

• 8_seed_story_synthesizer

  Use to enforce the final Research Seed Story Team output schema for orchestrator persistence, without adding rationale or process notes.

• 9_convener

  Fixed-voice pragmatic-chairperson persona for the Practical Research Narrative Team; responds to each attacker's challenge against Team 8's seed story with either a rebuttal grounded in the seed story or ground-truth inputs, or a revision of the challenged section, itself subject to evidence_discipl

• 9_it_engineer

  Fixed-voice hands-on-practitioner persona for the Practical Research Narrative Team; attacks Team 8's seed story's method and experimentation sections for underestimated implementation burden or a precondition the repository profile does not actually satisfy.

• 9_narrator

  Fixed-voice clear-technical-writer persona for the Practical Research Narrative Team; invoked exactly once after all six section cycles close, to turn this run's working document and the full attack-defend transcript into the final practical research narrative document.

• 9_professor

  Fixed-voice senior-authority persona for the Practical Research Narrative Team; attacks weak theoretical grounding, unsupported leaps from background to goal, and drift toward 0_problem_framing's out_of_scope boundaries in Team 8's seed story, then confirms whether the convener's response actually r

• 9_scout

  Fixed-voice resourceful-field-researcher persona for the Practical Research Narrative Team; hunts benchmark leaderboards, GitHub repos, and HuggingFace model/dataset pages for practical evidence to challenge or help defend Team 8's seed story, scoped to evidence the current story needs - never a re-

• 9_student

  Fixed-voice naive-but-persistent persona for the Practical Research Narrative Team; asks the basic clarifying question an expert glosses over in Team 8's seed story's method and conclusion sections, catching blind spots professor's authority-based attack does not.

MCP servers (1)

• literature-search

  runs: uv run --directory ${CLAUDE_PLUGIN_ROOT}/mcp_servers/literature_search literature-search-mcp

Similar plugins