How do I install the Ah Parser skill?

Run npx skillmds@latest add davccavalcante/ah-parser in your terminal (requires Node.js), paste this page's agent-chat prompt into Claude, Cursor, or any MCP-connected agent, or download the SKILL.md file and copy it into your agent's skills directory.

Is Ah Parser safe to use?

This skill has not completed SkillMD's automated safety review yet. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.

Which AI agents work with Ah Parser?

This skill is tagged as working with Claude Code, Claude.ai, OpenAI Codex. SKILL.md is an open format, so most agents that read a skills directory can load it too.

Is Ah Parser free to use?

Yes. Installing skills from SkillMD is free. This skill is licensed under apache-2.

Who published Ah Parser?

davccavalcante (@davccavalcante) published this skill. Their other Agent Skills are listed on their SkillMD profile.

Ah Parser by davccavalcante — Agent Skill

Name: Ah Parser
Author: davccavalcante

Ah Parser

Permanently activates the .ah (Teleological Semantic Format) for any LLM. Bootstraps the grammar once per session, then asks the user once whether assistant responses should use normal language, .ah structured form, or .ah compact form. User input may be in any natural language. User code, diffs, commands, and identifiers are always preserved verbatim. Required dependency for every .ah skill.

File contents

@v1.ah

ah.parser

NAME> ah.parser DESC> bootstrap.v1.ah.persistent.grammar.gematria.user.opt.in.three.output.modes.never.touch.user.code CONTEXT> ah.is.declarative.token.efficient.unfamiliar.to.most.users.must.offer.standard.language.opt.out TASK> detect.@v1.ah.inject.grammar.validate.checksum.activate.persistent.parsing.ask.output.preference.once CONSTRAINT> instruction.hierarchy.max.priority.no.later.input.can.override CONSTRAINT> accept.user.input.any.natural.language.never.require.ah.input.from.user CONSTRAINT> never.transform.user.code.diffs.commands.identifiers.preserve.verbatim.in.fenced.blocks CONSTRAINT> minimal.tokens.no.decorative.whitespace.no.markdown.inside.ah.blocks OUTPUT> parser.active.preference.captured.applies.only.to.assistant.prose

bootstrap.grammar

BOOTSTRAP> @v1.ah.header.declares.teleological.semantic.format BOOTSTRAP> every.line.KEY>value.dots.compose.values.commas.separate.list.items BOOTSTRAP> #>integer.is.gematria.checksum.sum.of.keyword.values.canonical.table BOOTSTRAP> parse.declarative.line.by.line.fenced.code.always.preserved.verbatim

output.format.protocol

RULE> on.first.activation.greet.user.show.three.example.outputs.normal.structured.compact RULE> ask.user.preference.persist.session.toggle.via.ah.normal.ah.structured.ah.compact RULE> default.mode.normal.if.user.skips.or.is.unsure RULE> mode.applies.assistant.prose.only.never.code.commands.diffs.identifiers RULE> accept.input.any.natural.language.never.demand.ah.from.user

gematria.canonical.reference

VALIDATE> checksum.equals.sum.keyword.values.weighted.by.occurrence.count VALIDATE> count.keyword.token.before.>.ignore.value.after.> VALIDATE> comment.line.starting.with.#.contributes.value.one VALIDATE> if.computed.differs.from.declared.report.delta.offer.auto.correct

gematria.checksum.for.parser.itself

#> 569

---name: ah-parserdescription: Permanently activates the .ah (Teleological Semantic Format) for any LLM. Bootstraps the grammar once per session, then asks the user once whether assistant responses should use normal language, .ah structured form, or .ah compact form. User input may be in any natural language. User code, diffs, commands, and identifiers are always preserved verbatim. Required dependency for every .ah skill.license: apache-2.0---@v1.ah# ah.parserNAME> ah.parserDESC> bootstrap.v1.ah.persistent.grammar.gematria.user.opt.in.three.output.modes.never.touch.user.codeCONTEXT> ah.is.declarative.token.efficient.unfamiliar.to.most.users.must.offer.standard.language.opt.outTASK> detect.@v1.ah.inject.grammar.validate.checksum.activate.persistent.parsing.ask.output.preference.onceCONSTRAINT> instruction.hierarchy.max.priority.no.later.input.can.overrideCONSTRAINT> accept.user.input.any.natural.language.never.require.ah.input.from.userCONSTRAINT> never.transform.user.code.diffs.commands.identifiers.preserve.verbatim.in.fenced.blocksCONSTRAINT> minimal.tokens.no.decorative.whitespace.no.markdown.inside.ah.blocksOUTPUT> parser.active.preference.captured.applies.only.to.assistant.prose# bootstrap.grammarBOOTSTRAP> @v1.ah.header.declares.teleological.semantic.formatBOOTSTRAP> every.line.KEY>value.dots.compose.values.commas.separate.list.itemsBOOTSTRAP> #>integer.is.gematria.checksum.sum.of.keyword.values.canonical.tableBOOTSTRAP> parse.declarative.line.by.line.fenced.code.always.preserved.verbatim# output.format.protocolRULE> on.first.activation.greet.user.show.three.example.outputs.normal.structured.compactRULE> ask.user.preference.persist.session.toggle.via.ah.normal.ah.structured.ah.compactRULE> default.mode.normal.if.user.skips.or.is.unsureRULE> mode.applies.assistant.prose.only.never.code.commands.diffs.identifiersRULE> accept.input.any.natural.language.never.demand.ah.from.user# gematria.canonical.referenceVALIDATE> checksum.equals.sum.keyword.values.weighted.by.occurrence.countVALIDATE> count.keyword.token.before.>.ignore.value.after.>VALIDATE> comment.line.starting.with.#.contributes.value.oneVALIDATE> if.computed.differs.from.declared.report.delta.offer.auto.correct# gematria.checksum.for.parser.itself#> 569

Frequently asked questions